Bose Quantum
国家支持的光子量子扩张公司,产品动能真实,但经济性不透明
Bose Quantum 具备战略重要性和足够可信的技术证据,值得继续尽调;但公开披露对收入、轮次条款和估值太薄,无法在未披露价格上承保。
封面要素
公司概况
Bose Quantum 是一家位于北京的光子量子计算公司,成立于 2020 年 11 月,业务把专用硬件、云端访问、SDK 工具和制造野心绑在一起。公开来源核实了 100、550 和 1000 量子比特产品代际、与 China Mobile 相关的云平台、深圳工厂建设,以及 2026 年 3 月 CNY 10 亿 B 轮融资,但收入、利润率和精确投后估值仍未披露。
- 成立时间
- 2020-11-16
- 创始人
- Wen Kai
- 创立地点
- Beijing, China
- 总部
- Chaoyang District, Beijing, China
- 产品
- 面向优化和企业或科研应用负载的专用相干光子量子计算机、Wuyue/Hengshan 云平台、Kaiwu SDK,以及相关控制和制造基础设施。
- 客户
- 科研机构、电信和云合作伙伴、超算中心,以及 AI、生物医药、金融、物流、能源和通信领域的企业用户。
- 商业模式
- 可能混合硬件系统销售、云端任务访问、软件 / 开发者工具和战略项目部署,但具体收入模式没有公开披露。
- 阶段
- Series B / growth
- 融资情况
- 2026 年 3 月完成 CNY 10 亿 B 轮,此前在 2023 年完成 A 轮、2024 年完成 A+ 轮、2025 年完成 A++ 轮;精确投后估值和条款在公开来源中仍未核实。
执行摘要
主要优势
- 真实硬件、云和工具面已经浮出水面:公开来源验证了 100-qubit、550-qubit 和 1000-qubit 系统,以及 Wuyue/Kaiwu 生态资产。
- 强投资人支持最终落到 CNY 1 billion Series B 和深圳制造推进上。
- China Mobile 等具名部署,比纯实验室创业公司给出更强商业化信号。
- 面向中国的战略重要性,可能让 Bose Quantum 持续嵌在国内政策与产业生态里。
主要风险
- 收入、毛利率、烧钱速度、现金跑道和准确投后估值仍未披露。
- 即便资金充足,光子量子商业化的技术和商业不确定性仍高。
- 工厂建设和芯片线计划会在重复需求被公开证明前抬高资本强度。
- 政策和出口管制暴露,可能收窄供应商、资金池或国际 go-to-market 选择。
未决问题
- Series B 的准确投后估值、证券类型和清算优先权。
- 经审计收入、毛利率、现金烧钱速度和现金跑道。
- 18M+ 云使用量指标背后的付费客户续约和合同经济性。
- 工厂良率、供应商图谱和出口管制缓释方案。
目录
01公司概览
1.1 身份、总部与阶段
Bose Quantum,英文也使用 Boson Quantum Technology 或 QBoson 品牌,是一家北京私营光子量子计算公司,现行法定名称为 Beijing Boson Quantum Technology Co., Ltd.,公司类型为非上市股份有限公司。登记记录和官网把公司定位在北京市朝阳区,成立日期为 2020-11-16,经营地址在万红西街。业务模式不是纯软件栈:公司称自己是全栈光子量子平台提供商,覆盖专用量子计算硬件、云端访问、开发工具,以及面向企业和科研负载的应用部署。公开证据支持的当前阶段,是一家处于 B 轮、正在扩张的私营硬件公司,而不是成熟商业运营商。最强公开来源确认了激进的技术和制造野心,但没有披露经审计收入、经常性软件经济性或干净的投后估值,因此这些指标仍是明确的尽调缺口,而不是事实。[CO001, CO002, CO003, CO004, CO005, CO006]
Bose Quantum 如何把光量子硬件、云工具、合作伙伴和资本接成商业化闭环。
[CO003, CO005, CO008, CO018, CO021, CO030]1.2 创始人、管理层与关键人物依赖
公开的领导层叙事高度围绕创始人兼 CEO Wen Kai 博士。英文行业报道、公司网站和学术记录都指向他的 Stanford 博士背景、Google 量子业务经历,以及早期参与相干 Ising 机研究。官方团队页还把 COO Ma Yin、CTO Wei Hai 和首席科学家 Wang Chuan 列为推动产品化的核心高管。登记类来源增加了一层薄但有用的治理信息:Aiqicha 将 Ma Yin 标为副董事长,并列出 Li Huiling、Jiang Peixing、Ruan Dong 和 Wang Dan 在财务或董事岗位任职。这足以说明公司已经不只是单一创始人叙事,但还不足以核实独立董事会监督、股东控制、董事会委员会或决策权。关键人物集中度因此仍然重要:技术故事、外部品牌和投资人信号仍都经过 Wen Kai 和创始技术团队传导。[CO007, CO008, CO009, CO010, CO011, CO012]
| 人物 | 职务 | 背景 | 创始人-市场匹配 / 覆盖 | 关键人物依赖 |
|---|---|---|---|---|
| Wen Kai | 创始人兼 CEO | Stanford 博士;前 Google 量子负责人;发表过量子计算研究 | 技术型创始人-市场匹配极强,外部可信度高 | 高 |
| Ma Yin | 联合创始人、COO / 副董事长 | 官网和注册记录列名、偏运营的联合创始人 | 连接产品化和公司治理 | 中 |
| Wei Hai | CTO | 官网列名的技术负责人 | 支撑架构和执行深度 | 中 |
| Wang Chuan | 首席科学家 | 官网列名的科学负责人 | 锚定研究可信度和技术方向 | 中 |
| Li Huiling | 财务负责人 | Aiqicha 列为财务负责人 | 补上财务职能,但公开运营细节很少 | 低 |
| Jiang Peixing / Ruan Dong / Wang Dan 等人 | 董事 | 注册类来源列名 | 显示正式治理面比英文媒体报道更宽 | 低 |
领导层表结合官网团队页职务和 Aiqicha 治理记录;公开董事权利细节仍有限。
[CO007, CO008, CO009, CO010, CO011]1.3 融资历史、投资人与资本信号
Bose Quantum 自成立以来多次融资,TMTPost 将 2024 年末的 A+ 延伸轮描述为公司成立后的第七次融资事件。公开记录对已命名轮次最清楚,对完整累计股权表则没有同等透明度。2023 年 A 轮超过 RMB 1 亿,并与 China Mobile 和清华系资本相关。北京政府背景基金领投 2024 年 A+ 融资;2025 年 10 月 A++ 轮引入新一批战略和财务投资人,金额达数亿元人民币;2026 年 3 月 B 轮达到 CNY 10 亿(约 USD 1.45 亿),财团由国资背景或机构关联投资人主导。模式比精确累计融资额更重要:Bose Quantum 正被当作国家战略型深科技制造商融资,而不是轻资本软件创业公司。即便如此,公开来源仍没有完整披露轮次条款、清算优先权、老股转让、债务或精确持股比例,因此除头部融资额之外的资本充足性仍需要直接尽调。[CO014, CO015, CO016, CO017, CO018, CO019]
| 利益相关方 | 角色 | 控制 / 经济重要性 | 尽调问题 |
|---|---|---|---|
| Beijing Financial Holdings / ICBC Capital / CMBI / Shenzhen Investment Holdings 等资本方 | Series B 轮领投资本方 | 高:锚定 CNY 1B 轮和国家相关战略支持 | 确认持股比例、董事会权利和任何政策任务 |
| Beijing Chaoyang Shunxi / Addor Capital 及其他 Series B 轮跟投方 | 后续成长投资人 | 中高:为制造扩产提供辛迪加深度 | 确认领投 / 跟投拆分和按比例跟投权 |
| Huade Tech Innovation / Nanshan Strategic Emerging Investment 等资本方 | 2025 年 A++ 轮共同领投方 | 高:资助工厂扩张前阶段 | 确认融资是否附带工厂激励 |
| GF Xinde / Hunan Caixin Industrial Fund / Weide Information / QF Capital 等资本方 | A++ 轮参与方和重复支持者 | 中:来自财务资本和上市公司资本的战略验证 | 确认后续储备和治理权 |
| Beijing High-Precision and Cutting-edge Industry Development Fund 等资本方 | 2024 年 A+ 轮领投方 | 高:政府背景资本支持商业化推进 | 确认支持是战略性、财务性,还是混合型 |
| China Mobile Digital New Economy Fund / Tsinghua Holdings Capital 等资本方 | 2023 年 Series A 轮领投方 | 中高:早期国家冠军和电信链路 | 确认商业拉动与纯财务支持的比例 |
| China Mobile / National Supercomputing Center Chengdu / BGI 等合作方 | 战略客户或合作伙伴,而非股东 | 中:需求验证和应用场景信号 | 厘清付费部署、合同金额和续约条款 |
投资人地图聚焦 2023-2026 年各轮公开点名的利益相关方;股权比例、清算优先权和老股转让不公开。
[CO016, CO017, CO018, CO019, CO020, CO021]1.4 封面指标、规模信号与披露缺口
可核实的头部指标偏运营而非财务。登记和公司来源支持 2020 年成立、现总部在北京、Aiqicha 报告参保员工 96 人,以及公司自称 70% 为研发人员、65% 为硕士或博士人才。产品使用证据比收入证据更强:TMTPost 报道 550 量子比特云平台累计超过 1800 万次任务调用,服务 830 个学科的 400 多家机构,并产出 440 多份应用报告;公司还声称 Kaiwu 社区拥有数千名开发者。这些指标说明生态触达,但不能证明付费客户数、经常性软件收入或硬件毛利率。精确投后估值在公开来源中同样不透明:订阅数据库追踪公司及其 B 轮状态,但开放访问输出无法独立核实被反复提及的独角兽估值。因此,我们把估值、收入、ARR、烧钱速度和跑道期都保留为没有公开支撑的封面指标。[CO006, CO013, CO025, CO026, CO027, CO030]
| 指标 | 数值 / 状态 | 截至 | 信心 | 缺口 / 备注 |
|---|---|---|---|---|
| 成立 | 2020-11-16 | 2026-07 | 高 | 注册信息和公司来源支持 |
| 总部 | 北京市朝阳区 | 2026-07 | 高 | 申报文件和官网地址一致 |
| 当前法律形式 | 其他股份有限公司(非上市) | 2026-06 | 高 | QCC/Aiqicha 显示 2026 年 6 月由 LLC 转为股份公司 |
| 参保员工 | 96 | 2026-06 | 中 | Aiqicha 参保人数;不是完整组织架构 |
| 研发人员占比 | 70% | 2026-07 | 中 | 公司关于页面口径 |
| 高级学位占比 | 65% | 2026-07 | 中 | 公司关于页面口径 |
| 最新披露融资 | CNY 1 billion B 轮 | 2026-03 | 高 | Xinhua、YiCai、QCR 和 TQI 广泛交叉印证 |
| 开源估值 | 2026-07 | 低 | 独角兽级估值没有在开源资料中独立验证 | |
| 收入 / ARR / 烧钱 | 2026-07 | 低 | 没有经审计的公开财务披露 | |
| 云采用标记 | 18M+ 次任务调用;400+ 家机构;830 个学科 | 2024-2025 | 中 | 使用指标,不是已确认付费客户 |
各行区分已验证经营事实和缺少支持的财务指标;null 表示本次查阅的开源资料没有验证该数值。
[CO001, CO002, CO003, CO011, CO012, CO013]头部成熟度与披露指标显示技术动能,但财务透明度仍不完整。
[CO001, CO013, CO017, CO025, CO026, CO030]1.5 里程碑、商业化与反向背景
公司时间线显示,它从科研创业公司快速转向制造主张方。Bose Quantum 成立于 2020 年末,2023 年公开发布 100 量子比特相干光子系统,同年晚些时候开通与 China Mobile 相关的 Hengshan 云平台;2024 年扩展到 550 量子比特机器并完成首个商用光子系统销售;2025 年又把 A++ 轮融资与深圳工厂建设放在一起推进。到 2026 年 3 月至 5 月,公司完成 CNY 10 亿 B 轮,发布 Yuliang Shanhai 1000 代际,并把自己定位为中国首家大规模专用光子量子制造商。时间线方向很强,但反向背景同样重要。Physics World 记录 Alibaba 和 Baidu 在 2023-2024 年退出直接量子硬件研究,说明即便资金更充裕的技术平台,中国商业量子市场仍然难做。因此,Bose Quantum 的里程碑速度值得关注,但还不等于需求已经去风险化,也不等于经济性已经透明。[CO014, CO017, CO018, CO020, CO030, CO031]
| 日期 | 事件 | 类型 | 金额 / 估值 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2020-11-16 | 公司在北京成立 | 创立 | 已注册 | Wen Kai 和创始团队 | 光子量子商业化努力起点 |
| 2023-03 | 完成 Series A 轮 | 融资 | >RMB 100M (~$14.4M) | 投资方:China Mobile Digital New Economy Fund;Tsinghua Holdings Capital | 硬件研发获得早期战略资本 |
| 2023-05 | 发布 100 量子比特相干光量子系统 | 产品 | 国内首台 100 量子比特相干光量子机器 | Bose Quantum | 证明硬件已越过概念阶段 |
| 2023-12 | 与 China Mobile 推出 Hengshan / Wuyue 云平台公测 | 合作 | 云平台上线 | China Mobile + Bose Quantum | 云访问扩大开发者和机构触达 |
| 2024 | 中国首台商业光量子计算机销售并开具税票 | 规模 | 完成商业交付 | Bose Quantum + 客户未完全披露 | 首次真实硬件货币化尝试的证据 |
| 2024-12 | 北京光电产业基金领投 A+ 延展轮 | 融资 | 数千万元人民币 | BOE/NAURA 支持的国家相关基金 | 强化商业化资本 |
| 2025-08 | 深圳光量子计算机工厂开工 | 规模 | 年产数十台目标 | QBoson / Shenzhen Nanshan | 制造野心从计划进入设施建设 |
| 2025-10 | 完成 Series A++ 轮 | 融资 | 数亿元人民币 | Huade;Nanshan;GF Xinde 等 | 支持芯片、工厂和通用路线图 |
| 2026-03 | 完成 Series B 轮 | 融资 | CNY 1B (~$145M) | 国家相关机构辛迪加 | 大规模制造和芯片试验线资金 |
| 2026-05 | 发布 Yuliang Shanhai 1000 和首款通用光子芯片 | 产品 | 1000 个专用量子比特;超频模式最高 3000 个 | Bose Quantum | 释放下一代产品和更宽平台野心信号 |
这是本次审阅中,Bose Quantum 关于创立、融资、产品、合作和规模里程碑的唯一有日期基准的记录年表。
[CO001, CO014, CO018, CO019, CO020, CO021]2020 年以来的创立、硬件、云、工厂和融资节点,一直到 2026 年产品与资本台阶抬升。
[CO001, CO014, CO018, CO019, CO020, CO030]1.6 证据要点
02市场分析
2.1 市场边界、纳入支出与现状替代方案
Bose Quantum 的相关市场是量子计算系统和服务,不是整个量子技术大伞。Bose 自身产品材料和 China Mobile 发布报道描述的栈,包含专用光子硬件、云端运行时访问、Kaiwu 开发者工具,以及面向应用的支持。QED-C 和 USCC 通过清晰区分量子计算、量子通信和量子传感,帮助界定这一市场。因此,纳入的支出池是硬件系统、云端或托管访问、工作流工具,以及面向仿真或优化问题的项目制应用工作。被排除或仅相邻的支出池包括通用半导体支出、广义 AI 基础设施、量子网络和传感预算,除非买方明确在资助混合量子工作流。这个区分很关键,因为 Bose 并不是在和所有企业计算竞争。它是在一组更窄的任务上,与经典 HPC、AI 模型和启发式优化工具竞争;这些任务的买方相信,光子量子最终可能在速度、搜索广度或仿真保真度上跑赢既有方法。[CM001, CM002, CM003, CM004, CM005]
| 细分 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 与 Bose Quantum 的相关性 |
|---|---|---|---|---|
| 专用量子计算硬件 | 光子或其他量子处理系统、控制模块、集成和安装 | 经典服务器、通用芯片、无关光学 | 国家实验室、HPC 中心、企业 R&D 或战略项目 | Bose 的核心切入点,因为产品页围绕专用光子机器展开 |
| 量子云访问 | 托管运行时访问、任务提交、监控、认证和托管使用 | 不含量子运行时的通用 IaaS | 研究机构、政府或企业创新团队 | 匹配 China Mobile 平台发布,并降低首批用户进入摩擦 |
| SDK 和工作流工具 | 开发者库、仿真环境、笔记本环境、模型转换和验证 | 非量子专用的通用 Python 工具 | 开发者、数据科学家、量子团队 | 支撑 Bose 的采用路径,因为 Kaiwu 是服务栈一部分 |
| 应用与共同开发服务 | 用例设计、试点支持、算法调优、垂直领域集成 | 不绑定量子工作流的广义咨询 | 创新、R&D 或业务单元发起人 | 可能是把实验性访问转成行业试点的必要环节 |
| 相邻但排除的类别 | 除上述量子计算栈外无其他纳入项 | 量子传感、量子通信、通用 AI/HPC 预算、广义半导体资本开支 | 买方和预算不同 | 防止 Bose 把全部量子或全部计算都纳入可服务市场 |
纳入行定义与 Bose 公开产品和云栈直接对应的支出池;排除行是相邻技术或通用计算预算,若计入会抬高 TAM。
[CM001, CM002, CM003, CM004, CM013]2.2 受证据约束的规模测算与相互矛盾的市场视角
公开来源支持市场分析,但不支持把 TAM 简化成一条线。QED-C 给出最干净的当前变现视角:2025 年量子技术市场为 $1.9B,其中计算板块为 $1.4B,并将在 2028 年扩大到 $3B 以上。McKinsey 提供另一种视角:已有 300 多家组织在与量子厂商合作,2025 年投资额飙升至 $12.6B。Post-Quantum 对 McKinsey 的总结把时间拉得更长,认为 2035 年内部量子技术市场大约为 $60B 至 $100B,其中量子计算贡献 $43B 至 $71B。这些数字不能直接相加,因为它们混合了当前收入、未来厂商市场预测和更广的经济价值池。对 Bose 来说,稳妥结论是:当前已变现需求的绝对规模仍小,但买方形成、资本形成和战略政策信号已经足以支撑一场真实的商业化竞赛。因此,最诚实的规模框架是受证据约束,而不是强行堆出 TAM-SAM-SOM。[CM006, CM007, CM008, CM009, CM010, CM011]
| 发布方 | 年份 | 地域 | 数值 / 视角 | CAGR | 方法 | 信心 | 局限 |
|---|---|---|---|---|---|---|---|
| QED-C | 2026 | 全球 | 2025 年量子技术市场:$1.9B;量子计算子板块:$1.4B | 行业层面年均增长 30% | 覆盖市场规模、投资、劳动力和 IP 的联盟记分卡 | 高 | 当前收入视角,不是 Bose 专属 SAM |
| QED-C | 2026 | 全球 | 量子计算预测:到 2028 年超过 $3B | N/A | 与 2026 年行业状态报告相关的前瞻市场预测 | 中 | 预测行只覆盖量子计算收入,不覆盖总经济价值 |
| McKinsey(经 Post-Quantum 摘要) | 2026 | 全球 | 2035 年内部量子技术市场:约 $60B-$100B | N/A | 第三方摘要的咨询市场模型 | 中 | 分母和时间范围不同于 QED-C 当前收入视角 |
| McKinsey(经 Post-Quantum 摘要) | 2026 | 全球 | 2035 年量子计算供应商市场:约 $43B-$71B | N/A | McKinsey 2035 年内部市场视角的子集 | 中 | 时间太远,不能直接换算成 Bose 份额 |
| McKinsey | 2026 | 全球 | 300+ 家协作组织 | N/A | 观察到的企业协作数量 | 高 | 买方池宽度信号,不是收入 |
| MERICS / USCC | 2024-2025 | 中国 | 约 $15B 的国家科研和产业支出视角 | N/A | 政策和生态分析 | 高 | 资金或基础设施视角,不是商业需求 |
各行有意混合当前收入、未来供应商市场预测、买方形成数量和国家资金视角,因为公开量子市场规模高度受边界影响;这些数字不应相加。
[CM006, CM007, CM008, CM009, CM010, CM011]证据受限的市场金字塔:从 2035 年广义经济价值,到 2035 年供应商市场视角,再到当前已货币化量子计算收入,单位均为十亿美元。
所有数值均按低端锚点以十亿美元计,便于各层共用单位;顶层是经济价值,中层是未来供应商市场视角,底层是当前收入,因此金字塔只表示方向,不能相加。
[CM012, CM033, CM041, CM042]量子市场不同视角的低—基准—高公开区间,单位均为十亿美元,显示今天的收入基数与 2028、2035 年展望的差异。
第 1 行覆盖 QED-C 对当前仅计算和全部量子技术收入的视角;第 2 行采用 QED-C 的 2028 年 >$3B 下限,以及 Post-Quantum 总结的 McKinsey 2028 年 $4.4B 视角;第 3 行采用 McKinsey 2035 年 $43B-$71B 计算市场区间,并推导中位数用于可视化。
[CM006, CM011, CM012, CM040]2.3 买方、用户和付款方分层;采用路径
公开来源中的买方故事,在科研和战略创新预算边缘最清楚,而不是在大众企业 IT 里最清楚。Bose 和 China Mobile 把今天的服务模式定位为云优先访问,面向科研用户、政府和企业团队,以及希望先通过 SDK 工具提交任务、再决定是否重部署的开发者。同业光子厂商也指向类似模式:用例集中在制药、化工、材料、金融、物流、能源和网络安全,而企业基准研究突出金融和汽车领域已经投入资源做量子实验的领先者。IQM 2026 年研究解释了为什么 Bose 短期买方很可能是成熟机构:89% 的受访者已经在动手实践,但只有 3% 达到规模化部署;预期使用混合或部分本地访问的买方,多于只使用公有云的买方。实践中,买方旅程仍是顺序推进,而不是一蹴而就。团队先学习,测试云端或混合工作流,共同开发商业用例,与现有系统集成,然后才考虑专用容量或重复生产采购。这个阶段式路径支持 Bose 的 SDK 加云切入点,但也意味着从生态使用转化为经常性付费收入,在公开来源中仍未被证明。[CM013, CM014, CM015, CM016, CM017, CM018]
| 细分 | 买方 | 用户 | 付款方 | 工作流 | 预算负责人 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 国家实验室 / HPC 中心 | 实验室主任或 HPC 项目负责人 | 研究人员、算法团队、领域科学家 | 政府或机构研究预算 | 云实验 -> 混合试点 -> 专用容量申请 | 研究项目办公室 | 需要在经典基线之外做仿真或优化 |
| 国家相关云 / 电信平台 | 平台所有者或战略技术单元 | 平台工程师和合作伙伴开发者 | 战略平台预算 | 将运行时接入托管云服务 -> 向用户开放任务访问 | 中央技术或创新预算 | 国家冠军定位和生态建设 |
| 制药 / 化学 / 材料 R&D | 计算科学或 R&D 负责人 | 研究科学家和建模团队 | R&D 预算 | 问题界定 → 算法试点 → 联合开发 → ROI 得到验证后进入生产 | 研发负责人 | 药物、材料或催化剂模拟受经典工具限制 |
| 金融机构 | 创新负责人或量化研究业务发起人 | 量化人员、风险团队、组合研究员 | 创新或量化研究预算 | 试点投资组合或风险工作流 → 混合集成 → 有限生产使用 | 量化研究或创新办公室 | 优化、风险或定价工作负载,能带来可衡量的延迟或质量收益 |
| 工业、物流和能源运营商 | 运营创新或数字化转型负责人 | 工艺工程师、规划人员、优化团队 | 业务单元转型预算 | 用例工作坊 → 云端试用 → 工作流集成 | 运营或转型预算 | 复杂路线、排程、能源或供应链优化痛点 |
买方、用户、付款方和预算负责人这些单元格,是从公开厂商页面、China Mobile 发布材料和企业采用研究推断出的、有证据支撑的原型,而不是披露的 Bose 合同条款。
[CM013, CM014, CM015, CM017, CM018, CM019]序数买方地图,显示 Bose 当前公开的光量子与云定位在哪些位置最强。
[CM020, CM031, CM036, CM037, CM043]观察到的量子采用路径从教育走向规模化部署,与 Bose 云 + SDK 的销售触点相匹配。
[CM013, CM017, CM018, CM034, CM035, CM044]2.4 增长驱动、采用约束与剩余尽调缺口
Bose 受益于真实的结构性驱动。中国国家主导的量子生态把大量战略资金、公私协调和实验室到市场的商业化压力导入该行业。QED-C、MERICS 和 USCC 都指向同一大格局:量子仍然资本密集、供应链敏感,也足够具有战略重要性,以至于政府和大型机构愿意在标准化企业需求显现之前补贴生态建设。这有利于 Bose 的本土市场。反向力量是,商业化风险仍高。人才稀缺、供应链脆弱、私营市场需求仍难观察,即便看多的来源也把这个市场描述为更奖励准备而非立刻规模化的市场。D-Wave 努力把自己的付费生产使用与行业其他玩家区分开来,反过来说明真正部署仍然罕见。尽调上,最大的剩余缺口都与 Bose 自身相关:公开来源没有披露定价、合同金额、从云端试用到付费生产的转化率,也没有透明的中国细分市场拆分。这些缺口让市场方向值得投资人研究,但使精确的 Bose SAM 或 SOM 仍无法落地。[CM021, CM022, CM023, CM024, CM025, CM026]
| 驱动因素 / 约束 | 方向 | 时点 | 影响 | 尽调问题 |
|---|---|---|---|---|
| 国家资金与公私协调 | 驱动 | 当前至 2030 年 | 加快生态建设,也让 Bose 在本土市场拥有比单靠私人需求更深的基础设施底座 | Bose 未来需求有多少依赖国家关联采购或补助项目? |
| 云端与 SDK 访问 | 驱动 | 当前 | 买方在投入专用系统前可以先学习和试点,漏斗顶部因此变宽 | Bose 云用户或开发者中,有多少会转化为付费试点或生产合同? |
| 光子模块化和可制造性主张 | 驱动 | 当前至 2035 年 | 可能提升可维护性、联网能力,也让最终本地部署更适合企业或 HPC 买方 | 哪些光子优势已经体现在实际采购胜单里,而不只是路线图表述? |
| 金融、医药和工业优化中的企业竞争压力 | 驱动 | 当前至 2030 年 | 即使完整容错尚未实现,也会催生实验预算 | 哪些垂直领域已经对 Bose 特定工作负载形成可重复付费意愿? |
| 跨学科人才短缺 | 约束 | 当前 | 拖慢试点设计、集成和买方内部准备 | Bose 有什么证据证明各垂直领域的客户成功能力和培训深度? |
| 先进量子组件供应链脆弱 | 约束 | 当前至 2030 年 | 增加硬件扩容执行风险,也可能拉长交付周期 | 哪些光学、控制和组件依赖仍在海外或单一来源? |
| ROI 证明和工作流集成要求 | 约束 | 当前 | 买方越来越看重校准访问、可解释性和集成,而不是量子比特数营销 | Bose 能否在具名客户问题上拿出经过基准测试的工作流收益? |
| 全行业规模化生产仍然少见 | 约束 | 当前至 2029 年 | 云端使用或试点未必能转化为持久经常性收入 | 有多少 Bose 部署已经付费、续约,并持续用于生产? |
驱动因素和约束混在一起,因为两者共同决定 Bose 表面的市场机会能否在风险投资时间表内变现。
[CM021, CM022, CM023, CM024, CM025, CM026]2.5 证据要点
03竞争者
3.1 直接的光子与架构同类公司
Bose 的公开材料明确显示,它卖的不是通用量子软件层,而是一套光子栈,覆盖专用硬件、室温运行、通过 China Mobile 的 Wuyue 平台云端暴露,以及 Kaiwu SDK 工具。这让它进入的买方讨论,更接近光子或光学优先厂商,而不只是超导或离子阱既有厂商。PsiQuantum 正沿硅光子路线推进实用级系统;ORCA 面向 HPC 集成销售机架式、室温 PT 系统;Quandela 提供云端和本地 MosaiQ 系统以及 Perceval 工具;Xanadu 继续把 Aurora 定位为可扩展、网络化、模块化的光子计算机。LightSolver 不是量子公司,但在战略上重要,因为它试图用模拟光学硬件而不是量子比特来赢下优化类预算。结果是,Bose 的核心模态有差异化,但并不独一无二。买方已经有多种方式押注光子、网络化光子或光学计算,而不必专门选择 Bose。因此,在已审阅材料中,Bose 最强的直接优势是本土渠道适配和部署叙事,而不是对室温或云可访问光子计算的独占主张。[CP001, CP002, CP003, CP004, CP005, CP006]
| 竞争对手 | 类别 | 规模 / 融资视角 | 目标客群 | 差异化 | 局限 |
|---|---|---|---|---|---|
| Bose Quantum | 直接光子参照 | 私营 Series B 硬件扩张阶段;公开收入未披露 | 需要硬件、云端和 SDK 访问的中国关联科研、企业和战略用户 | 室温光子系统、China Mobile 云集成、Kaiwu SDK、本土政策契合 | 无公开价目表、付费客户数量或广泛全球采购入口 |
| PsiQuantum | 直接光子同业 | 私营;定位公用事业级,已审页面无公开价目表 | 化学、材料、PDE、国防等高算力工作流 | 硅光子路线图、模块化架构、应用牵引的企业叙事 | 商业访问和定价仍不如既有云厂商透明 |
| ORCA Computing | 直接光子同业 | 私营;PT-1 和 PT-2 系统已有公开描述,具体收入未披露 | 需要机架式光子系统的 HPC、优化、AI 和科学用户 | 室温机架系统、混合工作流、按领域试点的 GTM | 已审页面更强调试点和架构,较少给出广泛市场采用指标 |
| Quandela | 直接光子同业 | 私营;公开披露 6-24 量子比特 MosaiQ 产品和 2,461 名活跃云用户 | 需要云端或本地光子访问的欧洲科研和企业团队 | 云端加本地交付、Perceval 工具包、灵活定价、能效型光子系统 | 相比 IBM 级生态,公开规模仍小;简单标价尚未完全标准化 |
| Xanadu | 直接光子同业 | 私营;已审 Aurora 新闻稿,保留来源无公开定价 | 关注模块化光子联网路径的开发者和机构 | 以联网化、模块化光子为定位,并有知名软件生态 | 已审保留来源给出的商业化细节少于 ORCA 或 Quandela |
| IBM Quantum | 全栈既有厂商 | 公开定价、本地部署选项和 300+ 成员网络 | 寻求可信托管访问的企业、学术和政府买方 | 已审样本中最透明的企业级打包,并掌握广泛生态 | 硬件路线不同;达到有意义规模仍需高价投入 |
| D-Wave | 退火 / 混合既有替代者 | 上市公司;2025 年收入 $24.6M,并提供公共云服务 | 重视生产工作流和混合求解器的优化密集型买方 | 公开收入证明、Leap 云、混合求解器打包、实时访问主张 | 架构更专用,并非直接光子等价物 |
所选集合覆盖买方最相关的直接光子同业、既有量子厂商,以及一个在保留的 2026 年公开材料中可见的架构特定替代者;它不是全球每一家量子创业公司的穷尽清单。
[CP001, CP003, CP005, CP006, CP007, CP008]证据支撑的序数地图,按公开商业化透明度(x 轴)和部署可达性 / 灵活性(y 轴)比较 Bose 与替代方案。
坐标轴是 1-5 的序数估计,只综合本次保留的公开证据。x 越高,公开价格、商业化指标和采购清晰度越可见;y 越高,云、本地或多环境部署灵活性越可见。
[CP012, CP014, CP016, CP017, CP023, CP024]3.2 既有厂商、云中介与替代路径
Bose 面临的更难竞争压力,可能来自光子之外。IBM、D-Wave、IonQ、Quantinuum 和 Rigetti 都已经把硬件、云端访问、开发者工具或商业化证明的某种组合,变成公开采购界面。IBM 发布免费、按量付费、项目制、高级和本地计划,并用 300 多成员网络强化这些方案。D-Wave 把退火和混合求解器包进云服务,宣称可用性,并在公开年报中展示真实收入。IonQ 强调与所有主流云和 SDK 集成,Quantinuum 则依靠保真度领先和 BMW、bp、SoftBank、Synopsys 等企业案例。Rigetti 的芯片到云叙事和可销售的 Novera 开发硬件展示了另一种既有厂商模式:让平台更容易试用、更容易教学,也更容易插入现有工作流。云中介会加深挑战。AWS Braket 和 Azure Quantum 让买方在同一界面比较不同硬件家族,使实验常态化,也削弱单一厂商锁定故事。NVIDIA 和 LightSolver 进一步扩大替代集合,主张部分优化或混合负载可以由加速经典系统或模拟系统赢下,而不必依赖纯量子硬件。[CP014, CP015, CP016, CP017, CP018, CP019]
| 购买标准 | Bose | ORCA | Quandela | IBM Quantum | D-Wave | 云经纪商(AWS/Azure) |
|---|---|---|---|---|---|---|
| 室温或轻数据中心部署叙事 | 强:官网强调室温运行,不需要真空低温设备 | 强:PT 系统为机架式且室温运行 | 强:光子系统按节能、可装入数据中心来销售 | 中:有本地部署和托管访问,但已审页面的硬件叙事并不以室温优先 | 中:云优先且支持本地部署,但卖点不是室温简化 | N/A:经纪层,不拥有硬件 |
| 公共云访问 | 中:China Mobile Wuyue 集成已公开 | 中:领域试点和 HPC 集成已公开;广泛公共云市场存在感较弱 | 强:专属云平台披露了活跃用户数和伙伴访问 | 强:平台、套餐和网络均公开 | 强:Leap 是核心 GTM 入口 | 强:专为多厂商云访问设计 |
| 本地部署或主权路径 | 未知至中:硬件存在,但公开合同结构未披露 | 强:PT 系统定位为可集成进现有 HPC 基础设施 | 强:MosaiQ 交付和合作伙伴数据中心托管已公开 | 强:有专门本地部署方案 | 中:有本地 QPU 访问,但云仍是主线叙事 | 弱:经纪商降低承诺成本,但不能替代厂商自有的主权部署 |
| 公开定价可见度 | 弱:未找到保留的 Bose 标价或合同费率 | 弱:已审页面未发布标价 | 中:灵活定价和预约机制已公开,但不是简单价目表 | 强:免费、按秒、年度和本地部署价格结构已发布 | 中:服务打包已公开,但已审页面没有分钟级标价 | 强:AWS 和 Azure 发布定价结构和服务商套餐细节 |
| 开发者工具与工作流抽象 | 中至强:Kaiwu SDK 已存在,但公开生态广度较不清晰 | 中:面向优化和 AI 用例的混合开发环境 | 强:Perceval、API 和 SDK 是云产品核心 | 强:Qiskit Runtime、Functions 和网络计划写得很明确 | 强:Leap 提供混合求解器和多层栈 | 强:经纪模式把跨厂商实验标准化 |
| 企业采购 / 信任界面 | 中国为中,全球较弱:China Mobile 集成有助本土市场,但公开的全球采购标记很少 | 中:技术可信度公开,但广泛采购证明有限 | 中:有伙伴背书和用户数,但足迹仍在形成 | 强:网络规模、套餐、支持和本地部署选项明确 | 强:有审计申报、公开收入和生产使用叙事 | 强:超大云厂商采购和市场入口降低采购摩擦 |
单元格仅概括截至 2026-07-01 运行日保留的公开证据。未知或弱项反映已审页面的可见度,不代表能力不可能或不存在。
[CP002, CP003, CP007, CP010, CP011, CP014]按类别比较直接光子厂商、现有大厂、经纪商和替代品在五个决策维度上的买方契合度矩阵。
强 / 中等 / 弱评级按竞争者类别概括本次保留的公开证据,而不是覆盖市场上的每个 SKU。目标是说明哪类替代方案最能匹配买方关切,而不是暗示同一类别里的能力完全相同。
[CP011, CP014, CP017, CP020, CP022, CP023]3.3 定价、分销与转换成本
定价透明度是 Bose 最明显落后于公开基准的地方之一。Bose 的产品和 SDK 界面展示硬件类别和工具,但不展示标价、消耗定价、最低合同额或任何已披露的付费客户单位经济性。相比之下,IBM 发布按分钟计费的方案,AWS Braket 解释 shot、task 和 reservation 计费,Azure Quantum 则展示 IonQ、Quantinuum、Rigetti 和 Pasqal 的合作伙伴专属方案。Quandela 也没有完整发布一张简单价目卡,但至少宣传灵活定价、预约服务和活跃云端使用。这种不对称很重要,因为买方不会只按架构做量子决策;他们比较的是访问模式、预算科目、采购摩擦、集成路径和可逆性。多平台并用越来越正常。买方可以用 AWS 或 Azure 测试多种模态,和 IBM 或 D-Wave 使用更成熟的托管访问,也可以停留在 NVIDIA 或 LightSolver 等替代栈上,同时继续评估量子。因此,转换成本看起来是中等,而不是绝对。只有当团队投入特定工作流工具、主权或本地部署,或长期企业项目时,转换成本才会升高。Bose 仍可能在中国胜出,因为渠道信任和本地对齐更重要;但离开这一语境后,不透明定价和较窄的公开分销证明会削弱锁定能力。[CP004, CP011, CP014, CP015, CP016, CP017]
| 方案 | 公开定价信号 | 计量单位 / 合同模式 | 包含能力 | 未知项 / 限制说明 | 影响 |
|---|---|---|---|---|---|
| Bose Quantum 硬件 + SDK | 保留的 Bose 产品或 SDK 页面未找到公开标价 | 未知;可能采用报价制硬件和服务合同 | 专用光子硬件、经由 China Mobile 的云端触达、Kaiwu SDK | 无公开单位定价、最低合同额或付费客户转化数据 | 外部买方更难对标总成本或采购速度 |
| IBM Quantum Open / PAYG / Flex / Premium / On-Prem 套餐 | 免费 Open Plan;PAYG 每分钟 $96;Flex 起步为 400 分钟/年、每分钟 $72;Premium 起步为 5200 分钟/年、每分钟 $48;本地部署按报价 | 免费试用、按秒计费、年度预购、年度订阅或专用系统 | 托管运行时访问、平台工具、支持和可选网络会员资格 | 达到有意义规模仍需要高价支出和合同 | IBM 为企业级量子打包树立透明度标杆 |
| D-Wave Leap | 公开服务打包、可用性和访问模式;保留页面无简单分钟费率表 | 通过云服务访问 QPU 和混合求解器 | 实时访问、生产级可靠性和优化工作流 | 保留页面上的定价细节不如 IBM 或 Azure 文档明确 | 由于服务和可用性主张公开,商业评估仍比 Bose 容易 |
| Amazon Braket | 按 shot 加按任务或按小时预约的定价公开 | 覆盖多种 QPU 和模拟器的用量制经纪模式 | 通过 Braket Direct 提供一致开发工具、混合工作流和专用预约 | 具体按 shot 价格因服务商和硬件类型而异 | 降低在厂商和技术路线之间切换实验的成本 |
| Azure Quantum marketplace 市场 | 市场和文档披露伙伴自定义套餐与定价结构 | 在同一经纪入口内提供按服务商划分的订阅、按量付费、token 或运行时计费 | 访问 IonQ、Quantinuum、Rigetti 和 Pasqal 产品 | 实际价格取决于服务商、workspace 和 Azure 基础设施费用 | 形成并排对标,而 Bose 尚未公开提供同等透明度 |
| Quandela Cloud | 灵活定价选项公开,但保留页面没有简单通用价目表 | 从小实验到企业使用的预约服务和平台访问 | QPU 访问、GPU 增强仿真、SDK/API 和伙伴托管选项 | 已审公开页面未保留具体费率表 | 访问模式比 Bose 更透明,但标准化程度不如 IBM 或 Azure |
| 通过 Azure 使用 Quantinuum | Standard 计划:每月 $125,000;Premium 计划:每月 $175,000 | 月度订阅,按 HQC/eHQC 统计用量,并有按量付费选项 | 通过 Azure Quantum 访问 H2 硬件和模拟器 | 需支付基础设施费用,工作负载成本仍取决于操作次数 | 展示既有厂商如何公开变现高端企业访问,而不是完全靠不透明报价 |
各行对比买方实际能查看的公开套餐和采购界面。有些方案披露精确费率,有些只披露结构或灵活性;这种差异本身就具有竞争意义。
[CP004, CP011, CP014, CP015, CP017, CP020]3.4 壁垒耐久性、信任姿态与反向证据
Bose 仍有真实资产。其官方产品页支持室温运行、可编程全连接、面向优化的性能主张,以及与 China Mobile 云生态集成。USCC 和 MIT 也支持一个判断:中国国家协调的量子建设,可以帮助本土公司获得政策关注、人才和早期基础设施伙伴。但这些政策动态也有两面性。MIT 认为,由于 QPU 多样性,美国在商业化上仍处于领先;USCC 则用国家战略视角描述中国量子崛起,这可能让跨境信任更复杂。公开证据也不支持对市场成熟度过度自信。QED-C 称整体市场仍只有 $1.9B,供应链脆弱、人才不足;D-Wave 备案文件显示,即便是行业内更商业化的厂商之一,仍担心销售周期加速和利润率压力。Physics World 又给出更尖锐的反向信号:Baidu 和 Alibaba 都退出了直接量子硬件研究。结论是,Bose 的壁垒更像有条件成立,而不是默认耐久。它可能在重视中国相关部署的本土战略客户中最强,但容易受到云中介比较、既有厂商生态力量、替代加速器,以及光子主张比客户信任更容易商品化的可能性冲击。[CP018, CP019, CP029, CP030, CP031, CP032]
| 护城河主张 | 威胁 | 严重性 | 缓释措施 / 尽调问题 |
|---|---|---|---|
| 室温光子硬件足以让 Bose 差异化 | ORCA 和 Quandela 也在营销室温或轻数据中心光子部署,Xanadu/PsiQuantum 仍保持光子规模化叙事 | 高 | 要求证明买方选择 Bose 不只是因为技术路线,例如赢单 / 输单记录、续约行为或垂直领域特定性能数据 |
| China Mobile 集成创造持久渠道锁定 | AWS、Azure、IBM 和 D-Wave 在其他市场把云中介实验和多厂商并用变成常态 | 高 | 要求 Bose 按国内战略客户与云试用用户拆分管线,并询问是否已有非中国云渠道上线 |
| 不透明定价有助于守住利润率 | 当 IBM、Azure 和 AWS 发布清晰访问结构时,不透明定价反而会拖慢采购 | 中 | 询问实际合同类型、最低交易规模,以及硬件销售、云访问和联合开发工作的定价是否不同 |
| 本土政策契合是纯粹的信任优势 | USCC 和 MIT 指出,这种在中国有帮助的国家关联姿态,也会让海外信任、出口和采购认知复杂化 | 高 | 要求出口管制评审、海外合规文件,以及中国以外任何公共部门采购限制 |
| 早期市场不成熟能保护领先者免受竞争 | QED-C 和 D-Wave 显示市场确有牵引力,但供应链仍脆弱、人才短缺、销售周期长,所有玩家都可能受罚 | 高 | 压测烧钱速度、制造韧性和从试点到付费生产的转化,不要假设品类增长会消化执行风险 |
| 优化工作负载天然属于量子厂商 | NVIDIA 和 LightSolver 说明,AI 超级计算机和模拟光学加速器不需要完整量子栈,也能吃到同一预算线 | 高 | 在认定护城河前,把每个 Bose 用例映射到真实的经典、模拟和退火替代方案 |
| 光子同业还太早,商业上无关紧要 | Quandela 已展示活跃云用户和交付时间线,ORCA 和 PsiQuantum 也在搭建企业应用叙事 | 中 | 按季度跟踪同业客户披露、伙伴数量和部署路径,检验 Bose 是领先者,还是只是拥挤光子队列中的一员 |
严重性评级是基于保留公开证据的分析师判断。登记表关注的是:当买方比较访问模式、信任界面和替代方案,而不只看量子比特标题时,Bose 能否守住差异化。
[CP029, CP030, CP031, CP032, CP033, CP034]紧凑的公开信号快照,显示市场围绕 Bose 及其替代方案披露了多少买方可评估的准备度。
指标混合了公开生态数量、披露收入和可见访问触点指标。它们不是标准化财务 KPI,而是基于公开来源揭示内容的买方可评估准备度视角。
[CP011, CP016, CP018, CP024, CP025, CP038]04财务
4.1 收入模式与定价界面
Bose 的商业界面比一份简单硬件手册更宽。公司公开展示的栈,从竞赛和开发者社区开始,经由 Kaiwu SDK 工具和云端验证,最终落到实体量子计算机部署。China Mobile 的 Hengshan 平台提供了一个特别有用的线索:注册后,用户可以在控制台订购真机算力服务,这让云端或按任务计费的收入路径变得可见,尽管 Bose 没有公布费率表。与此同时,Bose 自身页面没有展示硬件标价、按 shot 或按分钟计费标准,也没有最低合同信息。这迫使投资人用 IBM、AWS Braket 和 Azure Quantum 的透明同业产品来校准潜在买方预期。因此,证据支持一种混合收入模式——大报价硬件、使用驱动的云端访问、咨询式垂直解决方案工作——但无法清晰看见收入结构、实际定价、折扣或确认政策。公开定价不透明是真正的尽调问题,因为同业已经让量子访问采购清晰得多。[CI001, CI003, CI004, CI005, CI008, CI009]
| 收入流 | 机制 | 单位 | 当前价值 / 状态 | 质量信号 | 尽调问题 |
|---|---|---|---|---|---|
| 专用量子计算机销售 | 定制硬件系统销售和部署 | 按系统 / 项目 | 产品公开;有已交付客户报道;价格未披露 | 客单价高,但可能波动大且受验收驱动 | 平均售价、交付验收条款和硬件收入确认 |
| 云端真机访问 | 通过云控制台下单的任务式量子计算 | 按任务 / 计算订单 | 公开 beta 和下单流程可见;费率表未披露 | 最有希望形成经常性用量收入,但没有转化或留存数据 | 付费云用户、单任务价格和续约行为 |
| Kaiwu SDK 与开发者工具 | 与 Bose 硬件和云绑定的免费或捆绑软件入口 | 工具包 / 开发者席位 | 文档和 GitHub 公开;直接真机调用有记录 | 漏斗入口强,但变现可能间接 | SDK 访问是永久免费、捆绑,还是企业授权 |
| 垂直解决方案与集成工作 | 横跨 AI、生物医药、能源、金融和智慧城市用例的咨询式项目 | 按项目 / 试点 / 顾问费 | 多个解决方案页面有咨询 CTA;无合同金额 | 能制造战略证明点,但人力密集 | 服务构成、毛利率和各垂直领域可重复性 |
| 社区认证与开发者计划 | 免费机器配额、徽章和竞赛用于刺激采用 | 配额 / 促销额度 | 公开可见每年 10 个、每月 5 个 550 量子比特配额 | 有用的获客杠杆,但尚未证明可变现 | 配额到付费的转化,以及服务社区使用的成本 |
Bose 暴露了硬件、云、SDK 和解决方案入口,但公开来源未披露实际收入结构、合同金额或各收入流的确认政策。
[CI001, CI003, CI004, CI005, CI006, CI013]| 方案 | 单位 | 公开价格 / 替代指标 | 折扣或未知项 | 合同模式 | 来源 |
|---|---|---|---|---|---|
| Bose 硬件系统 | 按系统 | 无公开标价 | 按报价的企业销售 | 官方产品页面 | |
| Bose 云端真机服务 | 按任务 / 算力订单计费 | 未公开价目表或最低消费 | 注册后在云控制台下单 | China Mobile Hengshan 文章 | |
| Bose 开发者配额 | 机器配额 | 每年 10 次 + 每月 5 次 550-qubit 配额 | 促销性质,不是实际成交价格 | 社区认证与奖励 | Kaiwu 社区门户 |
| IBM Quantum 按量付费 | 按 QPU 分钟计费 | 96 | 年费方案费率更低 | 免费 / 按量付费 / 弹性 / 高级 / 本地部署 | IBM 官方定价 |
| AWS Braket | 按 shot + 按任务计费,或按小时预留 | 随 QPU 不同 | 按需或预留访问 | AWS 官方定价 | |
| Azure Quantum(IonQ 示例) | 按程序执行 / gate-shot 计费 | 97.5 | 关闭误差缓解后最低费用降至 12.4166 | 服务商定义的 PAYG 或订阅 | Microsoft Learn 定价 |
| D-Wave Leap | 询价 / 服务合同 | 抓取页面未公布商业定价 | 托管云访问,定位接近 SLA | D-Wave 云页面 |
同业价格只能作为可比采购代理,不能代表 Bose 实际成交价。Bose 在硬件和生产云两个公开界面上仍未披露价格。
[CI006, CI008, CI009, CI010, CI011, CI012]公开材料显示从开发者到云、再到硬件的桥接路径,但没有披露每一步的转化率或抽成率。
该流程只反映已发布漏斗和服务触点。Bose 未披露任何节点的转化率、货币化使用占比或收入权重。
[CI003, CI004, CI005, CI006, CI013, CI014]4.2 市场进入动作与销售效率代理指标
Bose 的公开 GTM 动作更像分层漏斗,而不是纯自上而下的资本设备销售。公司通过竞赛、社区参与、开放文档和 GitHub 示例吸引开发者,再用云端真机访问和额度激励降低试用摩擦。与此同时,三个垂直解决方案页面都把潜在客户导向咨询,这意味着有意义的变现很可能来自企业项目、试点或谈判合同,而不是标准自助结账。现有公开证据足以展示渠道形状,但不足以量化渠道经济性。我们没有看到披露的 CAC、回收期、NRR、胜率、市场转化率,或与 China Mobile 的收入分成条款。这意味着,可见的开发者和合作伙伴界面更适合作为管道代理指标,而不是收入质量证明。投资人应把社区、云端 beta 和咨询 CTA 视为严肃商业包装的证据,而不是 Bose 已经解决销售效率的证据。[CI003, CI004, CI006, CI007, CI025, CI026]
4.3 成本结构、单位经济性输入与资本强度
公开证据指向一家承担真实硬件和服务义务的公司。登记文件显示制造和云设备活动,关于页面则指向多个实验室和一座深圳工厂。产品材料补上了财务上重要的运营细节:日可用性目标、MTBF、MTTR、智能自检和功耗。这些指标意味着质保、现场服务、可用性工程和能源成本——正是让这门生意在结构上不同于纯软件公司的成本项。相比低温系统,Bose 的室温光子定位应当降低冷却和部署负担,但公司没有公布毛利率或服务利润率,因此这一好处无法转化为对单位经济性的信心。资本强度最强的独立确认,来自融资用途:芯片工艺能力、中试线建设、工厂扩张和生态建设。这是靠制造规模推进商业化,不是靠轻量软件铺开;这一区别是承销判断的核心。[CI016, CI017, CI018, CI019, CI020, CI021]
| 指标 | 公开值 / null | 置信度 | 为什么重要 | 尽调要求 |
|---|---|---|---|---|
| 收入 / ARR | None | 没有收入和 ARR,就无法检验商业规模或经常性质量 | 提供经审计收入、ARR 以及按收入流拆分的数据 | |
| 毛利率 | None | 毛利率是检验室温光子方案是否真正带来经济优势的核心 | 提供硬件、云和服务的毛利率 | |
| CAC / 回本周期 | None | 没有获客成本和回本周期,无法判断开发者漏斗经济性 | 按渠道提供 CAC、回本周期和销售周期数据 | |
| 净收入留存 | None | 如果云和 SDK 要形成经常性收入,留存就是关键 | 按队列提供 NRR 和 GRR | |
| 云到硬件转化 | None | 免费配额和云试用的价值,取决于能否转化为付费合同 | 提供从社区和云到付费部署的漏斗转化数据 | |
| 日服务可用时长(产品系列) | 6 到 16+ 小时/天 | 高 | 可用时长直接影响利用率和支持负担 | 按型号提供实际部署利用率和停机时间 |
| 功耗(已披露较大型系统) | 最高 1,200W 到 1,500W | 高 | 功耗会影响服务成本和部署占地 | 提供现场平均功耗和站点要求 |
| 可靠性 / 服务负担 | 已披露处 MTBF 500 到 1,000 小时;MTTR ≤72 小时 | 高 | 可靠性会影响现场服务人员配置和保修经济性 | 提供实际故障率、保修成本和备件政策 |
| 参保员工 | 96 | 中 | 员工数可粗略代理固定运营成本规模 | 按职能提供当前员工数和月度薪酬负担 |
公开单位经济性数字只剩运行技术代理指标,如可用时长、功耗、可靠性和员工数;实际收入和毛利率指标均未披露。
[CI019, CI020, CI021, CI028, CI029, CI036]公开单位经济性叙事基本是定性的:室温架构和自动化应该有帮助,但毛利率仍未披露。
之所以采用定性桥接,是因为 Bose 未披露服务成本、保修费用、利用率或已实现毛利率。
[CI019, CI020, CI021, CI043]公开证据指向制造、多地研发和服务交付的资本强度,但流动性披露仍缺失。
矩阵区分了公开证据已经支持的内容,以及没有内部数据就无法在财务上判断的内容。
[CI016, CI017, CI018, CI022, CI033, CI042]4.4 公开牵引代理指标与私有指标缺口
Bose 比许多深科技创业公司拥有更多公开商业化证明,但大多仍是代理证据。独立报道称深圳工厂已在 2025 年末投产,系统已经交付给 China Mobile 和国家超级计算成都中心等具名机构。China Mobile 文章还展示了一项向政府、企业和科研用户开放的云服务。在较软性的牵引端,Bose 声称拥有数千人社区,而 GitHub 和 Kaiwu 文档证明有真实的开发者上手路径。即便如此,真正决定性的数字指标仍然缺席:没有公开收入、ARR、毛利率、回收期、NRR、积压订单或客户集中度。结果是一个常见的量子投资问题:公司看起来商业活动很活跃,但公开数据无法说明这些活动是否转化为高质量经常性收入。我们可以框定部分技术成本输入,却无法看见足以支撑投资的实际经济性。[CI023, CI024, CI025, CI026, CI027, CI028]
| 缺失指标 | 对分析的影响 | 当前公开代理指标 | 为什么代理指标不足 | 精确尽调路径 |
|---|---|---|---|---|
| 按收入流拆分的收入 / ARR | 无法检验组合质量或集中度 | 已交付客户和公开云公测 | 商业活动不等于已确认的经常性收入 | 要求提供按硬件、云和服务拆分的经审计收入桥 |
| 实际成交价格和折扣 | 无法估算毛利率或销售效率 | 来自 IBM/AWS/Azure 的同业价目表;Bose 价格缺失 | 竞品资费不披露 Bose ASP、折扣或合同底价 | 要求提供价目表、交易审批折扣区间和前 20 大合同 |
| 按业务线拆分的毛利率 | 无法判断室温设计是否在经济性上更优 | 功耗、可用时长、MTBF 和自动化主张 | 技术输入无法披露服务成本或保修计提 | 要求提供硬件、云和服务的毛利率及队列趋势 |
| 社区到付费转化 | 无法把免费配额作为获客引擎估值 | 社区奖励和 GitHub 活动 | 使用激励可能带来互动,却不一定变现 | 要求提供从社区注册到付费算力再到硬件销售的漏斗数据 |
| 利用率 / 装机基数生产率 | 无法检验工厂或服务杠杆 | 工厂投产和对具名机构的部署 | 具名部署不披露已售机器小时或装机占用率 | 要求提供已部署系统利用率、积压订单和维护负载 |
| 现金 / 消耗 / 续航期 | 无法评估下一轮融资时点 | 融资新闻和注册资本 | 已融资金和注册资本不是流动性报表 | 要求提供现金余额、月度现金消耗、资本开支计划和现金续航期预测 |
| 债务 / 租赁 / 项目融资义务 | 无法评估隐藏固定费用或约束条款风险 | 未发现公开披露 | 沉默不等于零义务 | 要求提供债务明细、设备租赁和补助或补贴条件 |
| 客户集中度和留存 | 无法评估收入耐久性 | 具名机构交付和合作伙伴文章 | 标杆客户不显示重复支出或续约质量 | 要求提供客户集中度、流失、续约和按队列拆分的 NRR |
这里每一个缺失指标都对承销判断重要。公开牵引存在,但把牵引转成收入质量所需的私有运营数据并不存在。
[CI024, CI026, CI028, CI036, CI040, CI042]没有公开收入或利润率数据时,唯一有边界的量化输入来自已披露产品规格中的运营产能代理指标。
区间反映运营输入,不是财务输出。Bose 未披露公开收入、burn、runway 或利润率区间。
[CI020, CI021, CI029]4.5 资本充足性、融资依赖与结论
远期资本充足性是 Bose 的关键财务问题,公开证据只能部分回答。我们知道的是,公司反复融资,并把资金用于硬件研发、芯片工艺能力、工厂建设和生态扩张。仅 2026 年 B 轮就资助了芯片中试线和工厂扩产;登记数据还显示,2026 年 6 月公司改制为股份公司,注册资本增至 RMB 3.6 亿。我们不知道的是,资产负债表能否在不继续融资的情况下支撑下一阶段:公开来源没有披露现金、烧钱速度、跑道期、债务工具或项目融资承诺。D-Wave 的公开文件在这里是有用的行业基准:即便更商业化的量子玩家,相对于收入仍需要非常大的现金头寸。因此,Bose 的公开案例支持“商业化认真,但还不能靠公开信息承销”。正确的尽调姿态,是索要私有运营数据,而不是只从融资头条或工厂里程碑外推信心。[CI030, CI031, CI032, CI033, CI034, CI035]
| 资本项目 | 公开值 / 状态 | 来源视角 | 支撑什么 | 剩余缺口 |
|---|---|---|---|---|
| 注册资本 | 2026 年 6 月增资后 RMB360 million | QCC 登记 | 显示公司完成资本重整和组织形态升级 | 不是账上现金,也不披露现金续航期 |
| 实缴资本 | RMB17.98 million | QCC + Aiqicha 登记 | 显示法定实缴资本落后于注册资本 | 不披露当前不受限现金 |
| 2026 年 Series B 轮 | CNY1 billion | Yicai / TQI / QCR | 试点芯片线、工厂扩建、攻克技术瓶颈、量子 + AI 生态 | 融资后现金余额、现金消耗和资金投放节奏未公开 |
| 2025 年 Series A++ 轮 | 数亿元人民币 | 36Kr | 研发、芯片工艺能力、深圳工厂建设 / 运营、生态建设 | 确切金额和剩余流动性未公开 |
| 2023 年融资轮 | >CNY100 million | Yicai | 支持早期商业化,并绑定战略投资人 | 未公开融资后现金或里程碑约束条款数据 |
| 工厂状态 | 深圳工厂于 2025 年 11 月投产 | Yicai / TQI | 把资本转成制造产能,也带来库存风险敞口 | 未公开资本开支、利用率或营运资本数据 |
| 现金 / 消耗 / 续航期 | 无公开披露 | 决定公司是否很快需要下一轮融资 | 要求提供月度现金消耗、现金余额、现金续航期和资本开支计划 | |
| 债务 / 项目融资义务 | 无公开披露 | 可能实质改变稀释风险和约束条款压力 | 要求提供债务明细、设备租赁和任何项目融资 |
本表只用融资事实解释前瞻资本充足性。历史融资轮次在第 1 章;这里的关键结论是,硬件规模化让公司持续依赖融资。
[CI016, CI022, CI030, CI031, CI032, CI033]4.6 证据要点
05产品与技术
5.1 产品定义与模块图
Bose 的公开产品界面不止一台单用途机器。审阅过的官方页面显示,一套栈先从专用 100、550 和 1000 量子比特光子系统开始,再叠加云端访问、Kaiwu SDK 工具、面向 PyTorch 的插件、开发者激励和垂直解决方案模板。按工作流看,客户购买的不是抽象的“量子”,而是一条路径:表达优化或能量型学习问题,把问题路由进 Bose 的相干光子系统,并在 AI、生物医药和城市级场景中复用这条路径。因此,最强的产品定义信号是架构宽度,而不是无所不能。Bose 明确表示,其系统专门围绕 Ising 或 QUBO 式负载,软件栈则用来让现有 Python 用户保持生产力。这形成了连贯的产品故事,但也意味着差异化取决于这些相互连接的模块在生产中有多好用,而不只取决于量子比特头条。[CE001, CE002, CE003, CE004, CE016, CE017]
| 模块 / 资产 | 主要用户 | 公开角色 | 观察到的成熟度 | 关键差异化 | 尽调缺口 |
|---|---|---|---|---|---|
| SPQC-100 / 550 / 1000 硬件 | 科研、企业、战略买家 | 专用相干光子优化硬件 | 公开出货 / 有文档的产品系列 | 室温运行、全连接、多种机器规模 | 未公开按 SKU 拆分的验收测试结果或出货台数 |
| 量子云平台 / Quantum Cloud Hub | 开发者和云用户 | 通过托管云访问购买、排队并取回真机任务 | 公开界面在线 | 合作伙伴云触达,加上硬件池和故障切换叙事 | 未公开 SLA、状态页或经审计可用时长历史 |
| Kaiwu SDK | Python 开发者和研究人员 | 建模 QUBO 或 Ising 问题,并通过 Bose 工具提交负载 | 成熟文档界面 | 从问题建模直连真机 | 带凭证的 wheel 分发比标准 pip-only 流程更重 |
| Kaiwu-PyTorch-Plugin | AI 和应用研究团队 | 通过 PyTorch 暴露 Boltzmann-machine 和 QDiffusion 工作流 | 早期但具体的开发者层 | PyTorch 原生的量子采样叙事,而非通用 API 包装器 | 未公开采用指标、包注册表统计或 GitHub 以外的发布节奏 |
| Kaiwu 社区 / 示例仓库 | 学生、研究人员、开发者 | 示例、贡献入口、配额和认证漏斗 | 活跃引导层 | 动手示例和补贴的 550-qubit 使用量鼓励实验 | 从配额使用到付费生产的转化未披露 |
| 垂直解决方案(AI、生物医药、城市) | 领域团队和企业赞助方 | 把同一套优化栈改写成面向特定工作流的落地页 | 商业包装可见 | 场景框架降低买家从原始量子原语理解产品的负担 | 未公开深度案例研究,证明各垂直行业可重复的生产 ROI |
各行概括本章审阅的公开产品界面;它们描述 Bose 对外暴露的内容,而不是每个内部模块或控制台功能。
[CE001, CE002, CE003, CE008, CE016, CE017]| 用户任务 | 当前工作流 | Bose 方案路径 | 可衡量或声称的收益 | 观察到的限制 |
|---|---|---|---|---|
| 通用组合优化 | 在 Python 中建模 QUBO 或 Ising 问题,再用经典方法或专用硬件求解 | Kaiwu SDK 加云控制台或 CIMOptimizer 路径 | 同一问题表述可从本地建模迁移到真机执行 | 收益更像工作流便利性,而非独立基准优势 |
| 开发者在真机上实验 | 加入社区、赚取配额,并在购买硬件前尝试小型负载 | Kaiwu 社区测评加 550-qubit 配额 | 降低新开发者试用摩擦 | 配额激励不是付费留存证据 |
| AI 能量模型训练 | 在 PyTorch 中用 Kaiwu 后端训练 RBM、BM、QVAE 或 QDiffusion 工作流 | Kaiwu SDK 之上的 Kaiwu-PyTorch-Plugin | 让量子层更接近熟悉的 ML 工具 | 未公开生产客户案例量化插件带来的模型质量提升 |
| 生物医药或材料发现工作流 | 把序列、对接或设计任务映射为量子 + AI 优化问题 | 方案页面加插件 / SDK 栈 | 公开叙事强调更快搜索或更好探索大空间 | 公开页面仍是解决方案营销,不是端到端验证的客户研究 |
| 智慧城市优化工作流 | 把公交路线、能源调度、欺诈或波束成形转成优化任务 | 城市方案页面加云端或直接 SDK 执行 | 让产品锚定优化密集型运营问题 | 公开证据不显示采购条款、SLA 义务或长期生产经济性 |
工作流表停留在已有文档用户旅程和公开方案页面层面;不假设未披露的后处理或编排层。
[CE014, CE015, CE018, CE020, CE021, CE022]公开资料可见的路径:开发者定义问题之后,如何拿到 Bose 真机结果。
[CE010, CE014, CE015, CE018, CE020, CE021]5.2 架构、部署与可靠性
对一家私营量子创业公司来说,公开运营模型异常具体。Kaiwu 文档展示的开发者路径,从 Python 中的 QUBO 或 Ising 建模开始,经由求解器和优化器模块,再通过在云控制台上传矩阵或从代码调用 CIMOptimizer 触达真机。Bose 云页面随后补上编排层:统一硬件池、合作伙伴云分发、动态负载均衡和故障切换。硬件端,Bose 产品页强调室温相干光子运行、全连接、光纤温控和日益自动化的控制系统。这些并不等于独立证明,但足够具体,能看出公司认为运营价值来自哪里:减少低温负担、加强控制器集成,并围绕可用性提高自动化。技术风险在于,这些主张大多仍由公司自己撰写,因此可靠性证据作为架构叙事更强,作为第三方服务保证则弱得多。[CE004, CE005, CE006, CE007, CE008, CE009]
| 层 / 组件 | 在栈中的角色 | 关键依赖 | 观察到的风险 | 为什么重要 |
|---|---|---|---|---|
| 问题建模层(QUBO / Ising) | 把用户工作流转成可优化矩阵 | Kaiwu SDK 原语和求解器逻辑 | 开发者必须把问题塞进 Bose 接受的表述 | 这是入口,决定负载到底能不能被处理 |
| 求解器 / 优化器层 | 把模型送入经典、模拟或 CIM 支持的执行路径 | kaiwu solver、classical、cim、sampler、common 模块 | API 或求解器演进可能打断旧代码路径 | Bose 在这一层让非量子专家用上专用硬件 |
| 云端提交与队列管理 | 上传矩阵、配置任务、排队运行并取回结果 | 云控制台、账户凭证、任务编排 | 不透明的服务行为或队列延迟可能伤害用户信任 | 这是许多非硬件买家的公开生产界面 |
| Quantum Cloud Hub 资源池 | 把多台光子机器虚拟成一个托管资源层 | 硬件集群可用性和调度器健康度 | 故障切换和平衡目前只来自公司披露 | 资源池化是云页面上最清楚可见的运营护城河信号 |
| 机器控制系统 | 稳定运行模式、智能控制、诊断和环境监测 | L2 控制栈、传感器和自动化逻辑 | 缺少关于服务结果的独立证据 | 控制器质量看起来是可用性和支持成本的核心 |
| 光子计算核心和支撑模块 | 提供室温相干光子优化硬件 | 激光器、光路、光纤热控和全连接 | 硬件性能仍高度依赖 Bose 自己的描述 | 这是每个云或 SDK 工作流最终依赖的物理层 |
各行强调公开架构及其依赖,而非推测内部图纸。风险聚焦会破坏外部文档工作流的环节。
[CE004, CE005, CE009, CE010, CE011, CE012]从开发者工作流到光子硬件运营的六层公开架构。
[CE003, CE008, CE009, CE011, CE014, CE017]连接 Bose 软件、云、制造与硬件交付的关键公开依赖。
[CE008, CE009, CE012, CE028, CE029, CE030]5.3 差异化、研究谱系与制造
Bose 的差异化案例有三层。第一是模态选择:室温光子相干 Ising 硬件,瞄准优化式负载,而不是通用量子计算。第二是工具选择:面向 PyTorch、由示例驱动的界面,意在让开发者从熟悉工作流迁移到量子支持的采样或优化。第三是制造选择:比许多同业更早从实验室原型转向深圳工厂和芯片中试线。围绕相干 Ising 机和脉冲神经网络增强的研究谱系,有助于让架构更容易理解;但公开 IP 故事仍比公开工作流故事薄得多。同样,制造故事真实到足以重要——多个独立来源提到工厂、芯片线和质量控制功能——但仍缺少良率、吞吐量或出货单元证据。结果是,当前产品壁垒看起来是运营和工程牵引,还没有通过广泛独立基准或披露的单位经济性获得制度化证明。[CE017, CE018, CE024, CE025, CE026, CE027]
| 日期 / 阶段 | 里程碑或发布 | 公开状态 | 含义 | 来源 |
|---|---|---|---|---|
| 2022 | 文档描述 Kaiwu SDK 引用和软件包 | 历史版本,已发布 | 说明在较新的云和插件推进之前,Bose 已有具名开发者工具包 | Kaiwu SDK 介绍 / 引用区块 |
| v1.1.0 至 v1.1.1 | 新增直接求解器抽象、从模型到真机的教程、CIMOptimizer 和 PrecisionReducer | 已在文档更新日志发布 | 公开工具从建模辅助走向机器提交工作流 | Kaiwu 更新日志 |
| 2025 | Shenzhen 光量子计算机工厂开工,设模块、制造、质控 / 测试等部门 | 建设 / 部署阶段 | 制造正从原型叙事走向工业工程 | SCIO / Xinhua |
| 2025-2026 | 大额融资支持工厂运营和芯片中试线 | 扩产推进中 | 制造就绪度已进入产品论证,不只是资本开支 | Yicai / Quantum Computing Report |
| 当前旗舰代 | Shanhai 1000 具备三种模式、L2 控制和更长服务时长 | 当前商业旗舰 | 说明短期差异化落在专用系统的运营封装上 | Shanhai 1000 页面 / 规格 |
| 2026-2030 公开路线图 | 基于 Transformer 的调参、混沌幅度控制、多核并行和 CQ-H 光子芯片 | 愿景式 / 高层级 | 公开目标很大,但里程碑颗粒度低 | Bose 官网路线图 |
路线图表把已发布软件和工厂里程碑,与愿景式路线图拆开。后续里程碑只是公开方向信号,不是可用于投资判断的承诺。
[CE013, CE029, CE030, CE031, CE032, CE033]分析师矩阵:公开界面哪些部分看起来成熟,哪些仍缺文档。
评级概括了本轮保留的公开证据,并非 Bose 官方评分。成熟度更高不等于已获独立认证。
[CE013, CE017, CE029, CE033, CE034, CE037]5.4 信任、安全、合规与技术风险
公开信任界面明显弱于公开产品界面。Bose 确实发布硬件可靠性目标,提到 China Academy of Information and Communications Technology 验证,并在云端界面宣传企业级安全和数据隔离。但本章审阅的材料没有找到通常会让买方更容易承销这些主张的信任资料:没有保留下来的公开信任中心,没有明显的隐私或安全架构说明,没有公开证书页面,也没有可见的事故档案或状态历史。独立报道也提供了有用反向权重。Jiemian 记录的质疑认为,专用机器可能在商业上保持狭窄,且更快的经典硬件可能继续压缩付费意愿。这并不否定 Bose 的技术进展,但意味着投资人必须区分“有文档支撑的工作流成熟度”和“可信企业运营成熟度”。今天,前者可见;后者仍需要尽调资料包,而不是只读网页。[CE007, CE009, CE035, CE036, CE037, CE038]
| 控制或文档 | 公开状态 | 留存来源可见范围 | 支撑什么 | 缺口或注意点 |
|---|---|---|---|---|
| 硬件可靠性目标 | 发布在产品规格页 | Shanhai 1000 为 16+ 小时/天、MTBF >=1,000 小时、MTTR <=72 小时 | 表明 Bose 像运营商一样思考,而不只是实验室 | 未留存独立现场服务或可用时长日志 |
| CAICT 技术验证 | 产品规格页声称 | 可见一处技术验证提及 | 若能坐实,提供一定第三方质量信号 | 审阅页面未链接证书细节或测试范围 |
| 云池化和故障切换控制 | 云页面声称 | 24 台机器池、动态负载均衡、自动故障切换 | 表明围绕容量和韧性有运营工具 | 未公开 SLA 或事故档案验证生产表现 |
| 带凭证的平台访问和许可 | 文档和社区材料可见 | 平台登录、SDK 授权码、配额和任务提交控制 | 说明 Bose 通过自有平台和凭证控制机器访问 | 访问控制可见;数据治理和隐私控制不可见 |
| 公开认证和信任中心界面 | 保留来源未找到 | IAF 证书查询可作为验证路径,但保留来源没有留下 Bose 专属的公开证书材料 | 会影响企业采购和安全审查 | 信任中心、隐私、SOC、ISO 和状态页材料仍披露不足 |
本表把 Bose 已实际披露的内容,与成熟买家仍需要的信任材料拆开。这里的缺口只反映公开来源可见度,不证明内部控制不存在。
[CE007, CE009, CE035, CE036, CE037, CE040]5.5 证据要点
06客户
6.1 客户分层与访问界面
Bose 现在更像一家分层的量子基础设施厂商,而不是单用途硬件实验室。官方云和解决方案页面反复指向生物医药、金融、AI、交通或城市运营、材料、能源和科研用户,同时也区分谁只是基于这套栈开发,谁能够实际购买算力或硬件。最强商业信号不是宽泛的企业 logo 墙,而是明确访问模式、任务定价、私有部署表述,以及一条从开源或文档界面进入付费算力的真机软件路径。即便如此,界面仍不均衡。最宽的漏斗顶部似乎是高校、开发者和科研机构,而企业级证明集中在较小一组云、超算、金融和领域应用引用上。这让分层变得清楚,但还不能证明这些细分市场拥有同等变现质量。[CU001, CU002, CU003, CU004, CU005, CU006]
| 客群 | 买方 / 用户 / 付款方 | 可见证据 | 访问入口 | 生产级证明质量 | 备注 |
|---|---|---|---|---|---|
| 研究型高校和实验室 | 研究用户 / PI 或实验室 / 科研经费或机构预算 | 云页面列出 50+ 所高校和企业;高校仍占可见用户大块 | 开源 Kaiwu、云端按次付费、竞赛、学术共享计划 | 低-中 | 漏斗顶部可见度强,但公开续约数据弱。 |
| 生物医药和健康研发 | 科学家或平台团队 / 研究人员 / 实验室、药企或项目预算 | Jiemian 点名 Guangzhou National Laboratory、XtalPi 和 BGI;QbitAI 列出医院和高校合作 | 云任务、模型训练工作流、协作式应用探索 | 中 | 基础设施之外披露最充分的垂直领域,但大部分公开证据仍偏协作或工作流。 |
| 金融机构 | 创新或优化团队 / 量化或数据团队 / 银行科技预算 | China Merchants Bank 采购中标;Ping An 分支机构实地调研;更早的伙伴名单包括 Ping An 和 Huaxia | 按任务调用真机服务和定制优化支持 | 中 | 有具名证据,但长期扩张和重复支出未披露。 |
| 超算和公共算力中心 | 中心运营方 / HPC 和研究用户 / 公共或机构资金 | Chengdu 超算部署和金融科技任务测试 | 超算 + 量子一体化平台 | 中-高 | 这是最强的基础设施部署证据,但它可能仍偏探索,而非纯商业。 |
| 交通、城市和基础设施运营方 | 运营团队 / 规划方 / 企业或市政预算 | 官方城市页面和 Science and Technology Daily 点名 Shenzhen Metro | 解决方案项目和潜在私有化部署 | 低-中 | 具名证据弱于云、银行或超算。 |
| 开源开发者和学生 | 开发者 / 学习者 / 暂无直接付款方 | GitHub 仓库、ReadTheDocs、Kaiwu 社区、MathorCup 讲座 | 社区版、文档、竞赛和云登录 | 低 | 生态广度可见,但这更像用法探索,不是付费生产。 |
分层表区分可见用户和可见付款方;许多保留来源描述的是工作流或协作,不是合同经济性。
[CU001, CU002, CU007, CU010, CU013, CU014]| 指标 | 值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| China Mobile 公测启动 | Hengshan 光量子平台公测 | 2023-12-01 | QbitAI | 中 | 伙伴云分发早于大多数 2026 年融资头条。 | 启动时未公开付费用户数。 |
| 伙伴数量口径 | 50+ 所高校和企业 | n.d. | QBoson 云页面 | 中 | 说明其在高校和企业两端广泛铺关系。 | 未拆分活跃付款方、免费研究用户和沉默 logo。 |
| 云端调用量 | 6800w+ 累计求解调用 | 2025-10-15 | QbitAI | 中 | 这是量子平台的强活动信号。 | 调用量不等于独立付费账户或经常性收入。 |
| 机构覆盖 | 覆盖 900+ 家机构 | 2025-10-15 | QbitAI | 中 | 指向全国性的教育或研究覆盖。 | 机构覆盖不等于活跃生产部署。 |
| 开发者参与 | 10,000+ 名参与开发者 | 2025-10-15 | QbitAI | 中 | 说明开发者漏斗和反馈回路真实存在。 | 未披露开发者向付费企业用户的转化。 |
| 云端机器规模 | China Mobile 平台上的 100 个计算量子比特 | 2023-12-01 | QbitAI | 中 | 公有云访问绑定具体机器规模,不只是仿真营销。 | 未公开利用率或占用率。 |
| 用量驱动定价入口 | 每任务 128 RMB / 68 RMB 的列表价 | n.d. | QBoson 云页面 | 中 | 可见商业入口是交易型,试用门槛低。 | 未披露折扣、合同或企业打包。 |
轨迹表混合了公司声称和第三方报道的采用信号;这些指标都未披露付费账户数、ARR 或续约行为。
[CU002, CU004, CU009, CU011, CU022, CU023]Bose 如何把潜在客户从开发者认知推向云试用、具名机构部署,以及潜在的私有规模扩张。
阶段由保留的公开客户证据综合而来,不代表 Bose 的 CRM 阶段或内部漏斗定义。
[CU003, CU004, CU005, CU007, CU009, CU015]6.2 具名客户证明与证据成熟度
保留下来的公开证明是真实的,但并非同一种证明。China Mobile 云分发、成都超算部署和 China Merchants Bank 采购中标,是最清楚的具名界面:第三方可见,负载也比通用解决方案页面更具体。围绕这些锚点,Bose 还披露了跨生物医药、公共部门或交通场景的更深但更模糊合作。关键尽调区分在于,具名引用究竟证明的是已部署生产环境、付费试点,还是更接近应用探索的合作。公开来源目前把这些类别压在一起。这足以说明 Bose 已经摆脱纯概念阶段,但不足以假设每个具名伙伴都是大型、可续约、能贡献收入的账户。[CU009, CU011, CU012, CU015, CU016, CU017]
| 客户 / 机构 | 客群 | 部署或用例 | 生产 vs 试点 | 结果 / 证据质量 | 局限 |
|---|---|---|---|---|---|
| China Mobile Cloud / Hengshan 平台 | 伙伴云;政府 / 企业 / 研究用户 | 公测光量子云服务,配 Kaiwu SDK 和按任务调用真机访问 | 暗示可付费或可订阅的云访问;确切规模未披露 | 中等证明:具名伙伴平台、上线日期、用户路径和任务工作流公开 | 未披露该渠道的活跃账户数、留存或收入。 |
| National Supercomputing Center Chengdu | 国家支持的算力中心 | 550 量子比特部署,集成 100P 经典算力,并在金融科技工作负载上测试 | 已部署并测试,但收入模式未公开 | 中高证明:具名机构,加上 Xinhua 披露的部署细节 | 可能仍更多是验证基础设施,而非可重复商业需求。 |
| China Merchants Bank(Tiancheng AI)项目 | 金融服务买方 | 量子计算采购中标,包含按任务调用真机服务和定制优化支持 | 试点或早期生产型项目;期限未披露 | 中等证明:具名银行采购和服务范围公开 | 未公开合同金额、续约节奏或后续支出。 |
| 生物医药生态(Guangzhou National Laboratory、XtalPi、BGI、医院) | 研究和应用生命科学用户 | QBM-VAE 及相关工作流,覆盖肽、小分子、单细胞、疫苗和组学任务 | 从探索到试点,尚未明确披露为生产收入 | 中等证明:具名合作者和具名技术工作流 | 大多数证据偏协作或应用,而非商业。 |
| China Mobile、XtalPi、Shenzhen Metro 深度合作组 | 电信、生物技术、交通 | Science and Technology Daily 称 Bose 已在多个场景验证方案质量和效率 | 试点、部署和探索的组合不清楚 | 中低证明:权威媒体给出的具名关系 | 未披露项目经济性或持续时间。 |
列举范围仅覆盖本章保留的具名客户或伙伴证明;不是完整客户名单。
[CU009, CU010, CU011, CU013, CU015, CU016]| 界面 | 已明确证明 | 尚未证明 | 收入可见度 | 追加销售 / 扩张路径 |
|---|---|---|---|---|
| 开源仓库和文档 | 开发者可以建模 QUBO 问题、学习工作流,并看到真机路径 | 没有公开证据表明仓库参与会转化为付费账户 | 未公开 | 社区用户 -> 付费云任务 -> 企业项目。 |
| Kaiwu 社区和竞赛 | Bose 主动培育开发者和学生漏斗 | 未公开转化或留存 cohort | 未公开 | 竞赛参与者 -> 开发者账户 -> 应用型工作负载。 |
| China Mobile 云渠道 | 第三方分发和可下单的按任务真机服务已经存在 | 未披露经常性账户或收入数据 | 低-中 | 云试用 -> 重复工作负载 -> 私有化部署或更大的渠道交易。 |
| 超算中心部署 | 真机部署和测试过的工作负载公开 | 未公开合同经济性或续约证据 | 低 | 技术验证 -> 更广泛的机构项目或政府支持扩张。 |
| China Merchants Bank 项目 | 具名金融买方和明确范围的优化服务公开 | 未公开期限、TCV 或后续预算 | 中 | 试点或首次采购 -> 更多工作流或其他银行部门。 |
| 生物医药合作 | 具名伙伴和技术工作流公开 | 未公开收入分成或合同结构 | 低 | 研究合作 -> 生产级研发工作流或私有化部署。 |
本表刻意聚焦商业化,避免重复证明质量矩阵图。
[CU007, CU018, CU022, CU026, CU027, CU036]不同具名客户界面的公开证明质量差异很大;部署可见,但留存可见度几乎处处偏弱。
单元格是基于保留来源的证据判断,不是供应商认证的客户成功指标。
[CU007, CU009, CU015, CU018, CU021, CU023]6.3 耐久性、重复使用与仍未证明的部分
Bose 披露的使用活动已经足以说明人们在大规模接触平台,但公开耐久性记录仍然很薄。QbitAI 报道累计求解调用达到数千万次、覆盖数百家机构、开发者基数达五位数;云和文档界面也显示了明确的使用驱动工作流。尤其对新兴量子栈来说,这些都是有意义的采用信号。但在与续约、流失、NRR 或账户扩张数据绑定之前,它们仍只是漏斗上层和中层信号。最强的具名金融服务证明,是与 China Merchants Bank 的任务制合作;最强的公开云证明,是按次付费的合作伙伴平台。两者都更指向交易型使用,而不是长期订阅耐久性。除非 Bose 披露合同期限、重复支出或留存队列,否则外部投资人无法判断当前活动是有粘性,还是只是噪声很大。[CU004, CU007, CU008, CU018, CU019, CU022]
| 指标 | 值 / null | 客群 | 置信度 | 尽调追问 |
|---|---|---|---|---|
| 净收入留存 | 所有付费客户 | 低 | 索取按 cohort 和主要客群拆分的 NRR。 | |
| 总收入留存 | 所有付费客户 | 低 | 索取 GRR 以及按年拆分的 logo 留存。 | |
| Logo 流失 | 所有付费客户 | 低 | 索取账户赢单、流失及流失原因。 | |
| 合同续约率 | 云、银行、超算和合作 | 低 | 索取头部账户的续约时间表和续约结果。 | |
| 合同期限 | 具名银行和基础设施项目 | 低 | 索取具名部署的合同起止和延期条款。 | |
| 重复使用信号 | 68M+ 累计求解调用;10,000+ 名开发者 | 云和社区漏斗 | 中 | 拆分重复付费工作负载与免费或探索性活动。 |
| 使用模式信号 | 按任务定价和按任务提供的银行服务 | 云和金融 | 中 | 说明收入中一次性项目与经常性使用各占多少。 |
null 单元格不是漏译;它们反映保留来源缺少公开留存和续约披露。
[CU004, CU018, CU019, CU022, CU023, CU035]公开可见的推进路径:从广泛社区触达到更窄的付费或机构锚定部署界面。
该流程是定性的,因为公开来源披露的是各阶段示例,而不是每一阶段账户数全量。
[CU003, CU004, CU005, CU018, CU019, CU022]6.4 扩张循环与集中风险
Bose 确实有可信的扩张循环:社区和竞赛项目可以导入云试用,合作伙伴云访问可以导入具名机构部署,定制领域项目可以导入私有硬件或更广解决方案工作。问题在于,今天的公开证据仍集中在国资相关超算、合作伙伴云、科研浓度高的生物医药工作,以及少量金融引用上。Jiemian 的报道是这里最清楚的警示来源:它称超算中心是低风险首批客户,这类部署可能更偏探索而不是收入牵引,并且整个行业的重复客户经济性仍不清楚。Tencent 创始人访谈加入第二个约束,即当硬科技回报期很长时,银行和保险公司仍然谨慎。结果是,客户基础可信到值得关注,但仍集中在机构友好渠道;这些渠道可能比验证持久、多元收入更快地验证技术。缺失的桥是收入结构透明度:保留下来的公开来源没有量化需求中有多少来自云订阅、超算项目、定制银行工作或合作浓度高的科研账户,因此头部客户和头部渠道依赖仍是开放的承销问题,在承销集中风险前仍需要私有尽调。[CU014, CU020, CU021, CU031, CU032, CU033]
| 扩张驱动 / 风险 | 集中度风险 | 潜在影响 | 尽调路径 |
|---|---|---|---|
| 伙伴云和任务定价入口 | 中:试用容易带来大量用户,但不一定带来大量长期合同 | 表面活跃度可能跑在留存收入前面。 | 索取按渠道拆分的付费账户数、重复购买数和账户扩张历史。 |
| 超算中心首批客户模式 | 高:基础设施验证可能依赖国家关联或资助支持的机构 | 技术看似已被采用,但企业经济性可能还未跑通。 | 索取来自超算、公共部门和国家关联客户的收入占比。 |
| 生物医药合作深度 | 中:许多具名生命科学关系可能仍停在探索或项目制 | 强科学信号可能跑在重复商业支出前面。 | 索取探索伙伴转化为经常性收入客户的比例。 |
| 金融部门采购摩擦 | 中高:银行和保险公司回收期门槛高,采购谨慎 | 受监管行业放量可能慢于试点新闻暗示。 | 索取银行客户的采购周期、预算负责人和扩张里程碑。 |
| 开发者社区漏斗 | 中:如果云在生产中的价值不够强,大开发者基数也可能不付费 | 社区触达可能夸大商业化成熟度。 | 索取从社区注册到付费工作负载再到企业追加销售的转化漏斗。 |
| 经典方案替代压力 | 中:如果经典优化和 AI 硬件持续进步,专用量子需求可能收窄 | 通用量子准备好之前,续约和付费意愿可能先变弱。 | 索取相对经典替代方案的赢单 / 输单分析,以及未续约或试点停滞的原因。 |
风险表强调客户组合和商业化路径集中度,而非技术风险;技术风险见其他章节。
[CU031, CU032, CU033, CU034, CU038, CU041]6.5 证据要点
07风险
7.1 法律、政策与合规栈
理解 Bose Quantum 的法律风险,最干净的切口不是已知诉讼,而是监管边界正在外扩。已保留的一手和法律来源显示,美国对外投资规则已明确覆盖与中国相关的量子活动;围绕先进计算出口的 BIS 指引,也可能触达总部在中国的实体,即便跨境结构比直接发货更复杂。中国还叠加了自己的两用物项出口管制,以及互联网信息和类电信服务的本地审批。也就是说,Bose Quantum 同时暴露在边境两侧的监管里:境外投资人和供应商可能遇到尽调或许可摩擦,境内云服务或服务活动仍取决于本地许可和合规运转。国家优先方向是双刃剑。政策支持可以加快补贴、客户和产业协同,但也提高了一个风险:早期增长更像政策塑形,而不是市场已经验证。公开证据仍留下一个关键缺口:已保留来源没有独立排清所有名称变体下的诉讼、制裁或执法暴露。[CR001, CR002, CR003, CR004, CR005, CR006]
| 风险 | 法域 / 规则 | 为什么重要 | 可能性 | 严重性 | 缓释成熟度 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| 跨境投资者限制 | 美国对涉中国量子活动的对外投资规则 | 可能压缩合格投资者和 JV 范围,并在交易进入定价前增加尽调摩擦。 | 中 | 高 | 低 | 高 | 承销前,取得律师备忘录,梳理任何美国人士敞口、covered-foreign-person 状态和隔离结构。 |
| 高级计算出口管制摩擦 | 面向 D:5 总部实体的 BIS 指引和 EAR 管制 | 即便货物经第三国流转,也可能拖慢组件获取、工具流转或技术合作。 | 中 | 高 | 低 | 高 | 要求提供硬件、软件和服务触点的出口分类矩阵,以及任何供应商许可证被拒或延迟的记录。 |
| 中国两用物项出口管制合规 | 2024-12-01 生效的中国两用物项出口管制条例 | 可能限制技术、数据和服务跨境流动,也给先进系统增加境内合规义务。 | 中 | 高 | 低 | 高 | 审查内部出口管制政策、数据传输流程,以及外部合作或部署计划的法律顾问签批。 |
| 境内持牌服务暴露 | 互联网信息服务和增值电信许可 | 合规失误而非技术问题,也可能造成云端或服务中断,因为受监管业务已经进入商业触面。 | 中 | 中 | 中 | 中 | 核验现有许可证、续期时间表、具名合规负责人,以及 2025 年以来的任何监管通知。 |
| 诉讼 / 制裁可见度不完整 | 多个公司名称变体下的公开网页证据不完整 | 公开检索没有命中,不等于法律清查已经完成;制裁、执法或知识产权争议尤其如此。 | 低 | 中 | 低 | 中 | 交割前,要求按每个中英文名称变体做官方法院、执法、制裁和专利争议检索。 |
行按运行日已识别的最高严重度公开法律和监管暴露排序;这不是经法律顾问完整审阅的合规登记册。
[CR001, CR002, CR003, CR004, CR005, CR006]商业化、工厂执行与跨境合规相互强化之处,剩余敞口最高。
分数是从保留来源推导出的综合判断,用于相对排序,不代表精算概率。
[CR038, CR039, CR041, CR042]7.2 工厂扩产与技术瓶颈
运营风险高,因为 Bose Quantum 正在把仍未成熟的架构推向工业化。独立报道和公司材料都指向同一件事:公司用新融资扩建深圳工厂、搭建量子芯片试验线、提升可靠性;但这些材料没有给出良率、在线率或质保数据,无法证明转产已经受控。学术和行业证据也提醒谨慎。coherent-Ising 与光子计算文献表明,混合模拟-数字控制、光子集成电路、探测器、开关和其他光学层不是可互换的细节;每一层都是独立瓶颈,都可能卡住可制造性或性能。更广的供应链研究又加了一层风险:本土化确有进展,但还没补齐,专用矿物和零部件链条仍容易受地缘政治扰动。简言之,Bose Quantum 现在背着典型的「先建工厂、后证需求」风险:硬件野心更大、供应链变量更多,而公开证据仍不足以证明放大后的产品可靠到能支撑持久利润率。[CR010, CR011, CR012, CR013, CR014, CR015]
| 风险 | 失效模式 | 证据信号 | 可能性 | 严重性 | 缓释成熟度 | 剩余暴露 | 未解决缺口 |
|---|---|---|---|---|---|---|---|
| 工厂扩产快过良率爬坡 | 试点线或深圳工厂的人员与支出扩张快于稳定产出质量。 | 独立报道确认扩张计划,但未披露良率、质保或正常运行时间指标。 | 高 | 高 | 低 | 高 | 没有公开的良率、质保或 SLA 证据。 |
| 可靠性仍偏叙事 | 公司声称自动化和可靠性是目标,但运营历史尚未给出独立验证。 | QCR 和公司材料描述了可靠性目标;公开运营证据仍缺失。 | 中 | 高 | 低 | 高 | 没有公开事故档案或第三方可靠性审计。 |
| 混合控制系统复杂性 | 模数转换和反馈层限制速度、可维护性和可预测扩张。 | CIM 文献把混合控制描述为瓶颈,而非已解决的优势。 | 高 | 中 | 低 | 高 | 没有系统级披露故障率、校准负担或维护人力。 |
| 光子组件瓶颈 | 探测器、光源、开关和 PIC 都各自带来采购和可制造性风险。 | 行业图谱显示,光子堆栈上有多个相互独立的卡点。 | 中 | 高 | 低 | 高 | 未公开供应商图谱或合格二供清单。 |
| 矿物 / 材料暴露 | 铌、镍、稀土和其他特种材料可能带来隐性供应冲击。 | 独立供应链研究称,量子系统依赖集中度高且中国具备杠杆的材料。 | 中 | 中 | 低 | 中 | 未按 BOM 层级披露哪些材料对 QBoson 架构关键。 |
运营行把扩产、可靠性、架构和供应链失效模式拆开,便于管理层逐项证明缓释,而不是用一套泛化风险回应带过。
[CR010, CR011, CR012, CR013, CR014, CR015]最危险的失效路径,是出口或供应摩擦拖慢工厂产出,削弱部署证据,并触发又一轮融资需求。
边表示承销论点中的因果方向,不是实测弹性。
[CR014, CR017, CR019, CR041, CR042, CR043]7.3 合作伙伴、客户与模式脆弱性
Bose Quantum 的客户风险已经可信到必须重视,但公开面仍窄到令人担心。最强的公开证明来自合作伙伴云、超算部署和少数机构背书,而不是披露了大量续约的企业账户。Jiemian 的报道是关键反向视角:早期客户往往是超算中心,因为它们有公共资金、商业风险低;云端访问则通过 China Mobile、Alibaba Cloud 和 Huawei Cloud,以按次付费方式运行。这个分发模式可能帮助采用,但也意味着客户访问、认证和工作流控制有一部分留在合作伙伴基础设施里。披露出来的用户结构也偏集中,生物医药占比明显,大学仍在剩余部分中可见。财务模型风险直接来自这幅商业图景。收入、续约和大客户暴露都未披露,而公开量子同行即便披露更成熟,仍报告亏损、资本需求和客户集中。商业化因此是最大的剩余风险,而不是次要问题。[CR021, CR022, CR023, CR024, CR025, CR026]
| 依赖 | 交易对手 / 技术栈 | 角色 | 集中度信号 | 失效场景 | 严重性 | 缓释 | 剩余暴露 |
|---|---|---|---|---|---|---|---|
| 云分发合作伙伴 | 云渠道:China Mobile、Alibaba Cloud、Huawei Cloud | 客户访问、认证和按量付费交付 | 公开访问路径集中在少数具名平台上。 | 合作伙伴政策、定价或集成变化会压低访问量或利润率。 | 高 | 拓宽直销和本地化部署路径;写清合同保护。 | 高 |
| 机构锚定客户 | 超算中心和高校相关部署 | 验证与早期需求 | 可见案例偏向政策友好型机构,而非多元企业客户。 | 探索性部署未能转化为持久商业账户。 | 高 | 展示多年续约和非国有企业扩张。 | 高 |
| 生物医药需求集群 | 生物医药用户以及具名生命科学合作 | 使用集中度 | Jiemian 称,生物医药约占用户的 39%。 | 单一垂直领域停滞,或证明其不愿大规模付费。 | 中 | 披露按垂直领域划分的收入结构,以及向无关付费用例扩张的情况。 | 中 |
| 国资相关资本与政策生态 | 大型国资背景投资者联合体和国家优先级叙事 | 融资与生态准入 | 2026 年融资中能看到政策同向资本。 | 未来支持转向其他技术路线或采购优先项。 | 中 | 扩大商业客户基数,并加深民间资本厚度。 | 中 |
| 专用组件和工具链 | 光子供应商、PIC 制造、探测器和类 EDA 工具 | 制造就绪度 | 公开来源没有显示多元供应商图谱。 | 单一卡点会推迟出货或抬高商品成本。 | 高 | 认证二供,并在尽调中披露库存 / 交期纪律。 | 高 |
依赖行关注访问、收入和制造在哪些环节依赖外部平台或交易对手,而不是只靠 Bose Quantum 自身。
[CR021, CR022, CR023, CR024, CR025, CR026]| 角色 / 职能 | 依赖或缺口 | 证据信号 | 可能性 | 严重性 | 缓释 | 尽调路径 |
|---|---|---|---|---|---|---|
| 创始人与技术领导层 | 扩张仍显得高度依赖创始人和核心团队 | 公司叙事把创始人和技术领导层作为核心资产。 | 中 | 中 | 高 | 复核制造、产品和销售中创始人以下的授权深度。 |
| 制造运营与 QA | 工厂和试点线执行不能只靠研发深度 | 公开证据显示研发强度高,但 QA 人员配置细节有限。 | 高 | 高 | 中 | 要求提供生产工程、质量、现场服务和质保负责人的组织图。 |
| 合规与法务运营 | 跨境和境内受监管活动需要专门负责人 | 风险边界已覆盖出口管制、电信式许可,以及数据 / 服务流程。 | 中 | 高 | 低 | 索取具名合规负责人、外部法律顾问沟通节奏和升级日志。 |
| 法律形式变更期间的治理 | 2026 年,董事会和人员变化与资本和法律结构变化同时发生 | 登记更新显示人员和公司形式同步变化。 | 中 | 中 | 中 | 复核董事会会议纪要、授权事项和转制后的控制环境。 |
执行行强调,下一阶段可能掉链子的地方不是科学,而是组织没有搭到足够厚。
[CR034, CR035, CR036, CR037]可见商业化依赖一组紧密要素:合作伙伴云、机构锚点、政策资本和专用投入品。
该图展示依赖集中度,不展示所有权比例或合同规模。
[CR021, CR023, CR025, CR026, CR040, CR044]7.4 缓释准备、监测与退出标准
Bose Quantum 确实有实质缓释因素:技术型创始团队、很高的研发人员占比、近期资本,以及足够公开证据证明它不是纸面公司。但这些缓释因素没有补上最大的投资判断缺口。公司仍需要拿出私下证据,覆盖出口管制分类、供应商集中度、工厂良率和在线率、收入结构、续约,以及客户集中度。因此,正确的投资人姿态是有条件推进,而不是悲观否决。如果管理层能证明试验线按计划推进、云或硬件在线率有记录、合规人员配置可信,且至少少数客户在政策相关锚点之外续约或扩容,风险堆栈就更可控。如果这些材料不存在,今天让 Bose Quantum 显得有野心的同一组事实——工厂扩产、国家相关动能和专用硬件——就会变成放慢甚至退出的理由。否决标准应盯住延期、集中度和融资依赖,而不只看量子比特标题。[CR034, CR035, CR036, CR037, CR038, CR039]
| 风险 | 可监控触发器 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 对外投资 / 出口管制摩擦 | 法律顾问提示涵盖交易或供应商许可延误 | 出现任何禁止投资认定,或任何关键组件许可证延误超过 90 天 | 在依赖跨境资本或组件假设前,暂停承销或重构交易。 |
| 工厂与良率风险 | 试点线或工厂延期,且没有产出质量证明 | 延期两个季度,或到下一轮融资时仍没有良率 / 正常运行时间 / 质保材料包 | 下调扩张假设,把制造扩张视为资本消耗,而非护城河形成。 |
| 客户集中度风险 | 收入结构仍未披露,或集中在少数机构 / 单一垂直领域 | 最大垂直领域收入占比超过 40%,或前三大客户超过 60% 且无续约证据 | 下调商业化信心,并要求集中度契约或里程碑式出资。 |
| 合作伙伴平台依赖 | 云分发路径丧失或重新定价 | 失去主要合作伙伴云渠道,或平台经济性发生重大不利变化 | 重做增长模型,并要求提供直销或本地化替代路径的证据。 |
| 资本强度风险 | 收入耐久性得到证明前,又进入下一轮融资周期 | 本轮交割后资金续航少于 24 个月,或单位经济性没有改善路径 | 在耐久性得到证明前,不要按规模化收入倍数给公司定价。 |
| 运营证明缺口 | 未出现独立正常运行时间、可靠性或基准测试证据 | 没有第三方运营材料包、没有事故历史,也没有能证明生产耐久性的客户推荐 | 将业务视为尚未证明的基础设施,建议维持在继续研究 / 回避区间。 |
这些触发器旨在尽调或下一次刷新周期中可监控;每一项都把叙事风险转成具体决策阈值。
[CR038, CR041, CR042, CR043, CR044, CR045]08估值
8.1 建议、投资论点与反论点
Bose Quantum 的动能足够支持继续尽调,但披露不足以支持现在投资。可投论点很直接:公司显然不只是实验室项目,短时间内完成多轮融资,选择围绕光子专用量子系统搭建,而不是等待通用容错量子计算,并且有足够产品和部署证据,说明客户与国家相关伙伴认真看待它。反论点同样重要。公开来源仍未披露收入、利润率、轮次条款或投后估值;Jiemian 的反向报道说收入仍未披露,早期客户往往是超算中心,而不是广泛且续约密集的企业客户群。公开量子可比公司显示,投资人可以为巨大的期权价值支付高溢价,但这些公司披露更多财务信息,同时经济性仍波动很大。审慎结论是 **research-more**,信心为 **中等**,风险为 **高**;估值立场不是「明显有吸引力」,而是 **仅凭已披露事实仍未知**。[CV001, CV002, CV003, CV009, CV011, CV012]
| 维度 | 取值 | 当前落点原因 |
|---|---|---|
| 建议 | 继续研究 | 技术和融资证据足以保持跟进,但价格和财务披露不足以投入资本。 |
| 置信度 | 中 | 公开证据方向一致,但重大承销字段仍未披露。 |
| 风险评级 | 高 | 商业化、政策和融资不透明都很关键,而且彼此叠加,而不是相互分散。 |
| 估值立场 | unknown | 公开信息没有投后估值或条款清单;任何激进加价都需要私下收入和续约证据支撑。 |
| 决策含义 | 数据室证据到位前,不担任领投或定价角色 | 继续跟踪公司,但在承销轮次前必须看到硬尽调证据。 |
判断字段对价格敏感,因此刻意保守;保留下来的公开证据缺少估值条款和经审计的经营经济性。
[CV042, CV044, CV045, CV046, CV047]| 立场 | 论点 | 何种情况会改变判断 |
|---|---|---|
| 正方 | Bose Quantum 显然正在商业化真实的光子量子产品触面,而不只是发表研究。 | 如果产品主张不再转化为部署或客户案例,证明面会迅速变弱。 |
| 正方 | 2024-2026 年融资节奏显示,投资者反复有胃口,资本也足以建设制造能力。 | 如果下一轮规模小、仅由内部人出资,或结构上惩罚性强,今天的融资叙事支撑力会下降。 |
| 正方 | 早期销售、云端使用和专用定位,给出一条比通用量子路线图更早实现效用的可行路径。 | 如果这些早期部署不能转化为重复付费项目,专用捷径的价值会下降。 |
| 反方 | 收入、利润率和资金续航期仍未披露,投资者无法判断采用是真实需求,还是战略性试点活动。 | 经审计收入、客户分群和毛利率证据会显著提高置信度。 |
| 反方 | 公开融资文章披露了募资金额,但没有披露投后估值、证券类型或优先权条款。 | 完整股权结构表和条款清单,可把估值立场从未知转为可分析。 |
| 反方 | 跨境政策和出口管制不确定性提高了供应商和部分外国资本池的风险。 | 可信的出口管制备忘录和供应商图谱会降低政策折价。 |
行把可投资优势与最可能改变建议和定价纪律的证据缺口拆开。
[CV001, CV002, CV003, CV005, CV009, CV010]建议从真实技术与融资动能出发,落到尚未解决的财务和政策问题;这些问题卡住了定价。
流程是分析性而非统计性;它概括了建议背后的关键门槛逻辑。
[CV002, CV003, CV011, CV020, CV041, CV044]8.2 融资背景、定价纪律与证据边界
融资故事是档案里最清晰的利多事实,也是最需要保持纪律的理由。Bose Quantum 披露了 2024 年 A+ 轮、2025 年后续 A++ / A+ 轮第二阶段,以及 2026 年 3 月 CNY 1 billion 的 B 轮。B 轮规模足够大,投资人名单也强,但所有已保留公开来源都没有给出投资人真正需要的条款:没有披露投后估值,没有证券类型细节,没有清算顺位,也没有把融资规模与经营经济性接起来的公开桥梁。这一点很关键,因为公司把资本投向芯片试验线、深圳制造扩张和技术放大,而公开来源尚未证明收入质量可以重复。换句话说,Bose Quantum 在透明经济性出现之前,先为工业产能融资。这可能造出超额赢家,也可能留下稀释和优先权包袱,而公开文章看不见这些。任何投资判断都应把披露融资规模视为支持证据,而不是价格公允的证据。[CV003, CV004, CV005, CV006, CV007, CV008]
Bose Quantum 在动能上得分高,在金融级证据上得分低,因此结论仍是 research-more。
分数是 1-10 的分析师判断,只基于保留的公开证据;不是公司报告的 KPI。
[CV002, CV003, CV009, CV011, CV014, CV020]8.3 公开与私有可比样本
可比样本有信息量,但也危险。公开纯量子标的目前以异常高的市值 / 收入倍数交易:按最新可取得年度收入计算,IonQ 约 152.9x,D-Wave 约 361.0x,Rigetti 约 904.2x,Quantum Computing Inc. 约 3211.1x。这些数字说明市场愿意为量子期权价值出钱,但不能证明 Bose Quantum 应该得到同等待遇;恰恰相反,它们显示整个品类仍高度投机。私有样本同样跨度很大。Quantinuum 在 2025 年披露,以 $600 million 融资对应 $10 billion 投前估值;ORCA 更早的光子 Series A 只有 $15 million。因此,Bose 位于很不一样的可比坐标之间:比早期光子创业公司更有规模、融资更充足,但披露远不如公开纯量子公司,成熟度也远低于融资最足的私有龙头。可比公司只能提供框架,不能直接给出定价答案。[CV021, CV022, CV023, CV024, CV025, CV026]
| 可比公司 | 类型 | 最新披露收入或融资锚点 | 市值 / 估值锚点 | 隐含倍数或状态 | Bose 参考意义 | 主要局限 |
|---|---|---|---|---|---|---|
| IonQ | 上市量子平台 | $130.0M FY2025 收入 | $19.88B 市值 | ~152.9x 收入 | 展示最高质量的公开市场期权性基准,且有真实规模和现金。 | 披露、流动性和多元化程度都远高于 Bose Quantum。 |
| D-Wave | 上市退火 / 量子系统 | $24.6M FY2025 收入 | $8.88B 市值 | ~361.0x 收入 | 是仍在变现早期采用的上市量子公司的有用基准。 | 技术路线和客户结构不同;仍受益于公开市场流动性。 |
| Rigetti | 上市超导量子 | $7.1M FY2025 收入 | $6.42B 市值 | ~904.2x 收入 | 说明即便收入很小,量子期权性也能维持很高估值。 | 技术路线不同,且现金余额异常大,扭曲了直接比较。 |
| Quantum Computing Inc. | 上市光子 / 量子邻近 | $0.682M FY2025 收入(Stock Analysis);$0.373M FY2024 10-K 收入 | $2.19B 市值 | ~3211.1x 收入 | 最接近公开市场光子风格情绪标尺,但收入基数很小。 | 业务组合包括传感和光子学,超出 Bose Quantum 当前画像。 |
| Quantinuum | 非上市规模化量子龙头 | 2025 年 $600M 股权融资 | 投前估值 $10B | 私营后期龙头 | 规模和声誉强得多时,这一案例给出私营量子公司融资胃口的上沿。 | 不是中国光量子私营公司的可比样本,运营成熟度也高得多。 |
| ORCA Computing | 私营光量子初创公司 | 2022 年 $15M Series A | 仅早期融资 | 早期光量子参照 | 可参照早期私营投资人如何给光量子类别投钱。 | 阶段太早、规模太小,无法直接给 2026 年的 Bose 定价。 |
可比样本组混合了上市纯量子公司和私营参照点,因为没有一个已披露标的能同时匹配 Bose Quantum 的技术路径、地域、披露程度和阶段。
[CV021, CV023, CV024, CV026, CV027, CV030]承销区间中点受披露和商业化证明影响更大,单靠融资标题难以推动。
数值是以百万美元计的分析师敏感性标记,不是公司披露估值;每个条形围绕基准情境中点隔离一个门槛变量。
[CV015, CV016, CV043, CV048, CV049, CV050]8.4 牛 / 基准 / 熊情景与投资判断区间
Bose Quantum 不披露收入或估值,因此情景分析应被读作投资判断区间,而不是所谓公允价值。牛情景假设公司把当前战略牵引转化成可重复的企业或政府续约,证明试验线产出和良率,并把 2026 年融资变成可见商业化,而不只是扩产。在这个世界里,低于 $1.2 billion 的估值区间可以用私有市场期权价值逻辑来论证。基准情景保守得多:公司继续推进技术和融资,但公开证据仍以试点为主,下一轮融资更依赖故事质量,而不是经审计经济性;这支撑的是数亿美元中段区间,而不是跃上十亿美元。熊情景假设先建工厂、需求未跟上,出口管制摩擦,或下一轮疲弱且条款惩罚性强。沿着这条路径,价值会急剧压缩,普通股风险快速上升。区间的意义不是精确,而是说明定价敏感性主要由披露、重复需求和条款决定。[CV042, CV043, CV048, CV049, CV050]
| 情景 | 核心假设 | 示例价值区间(USD) | 概率信号 | 下行 / 上行驱动因素 |
|---|---|---|---|---|
| 牛市 | 重复客户、披露收入规模、试点线证明,以及可控政策摩擦 | $0.8B-$1.2B | 需要从战略性证明硬转为商业证明 | 上行来自稀缺光子量子期权与已披露经济性相匹配 |
| 基准 | 技术进展继续,但商业化仍以试点为主,下一轮融资仍依赖叙事加战略资本 | $0.35B-$0.6B | 与当前公开证据最一致 | 价值受收入、条款和续约质量不透明拖累 |
| 熊市 | 工厂扩张快过需求、政策摩擦恶化,或下一笔资本以过桥 / 降估值轮形式到来 | $0.1B-$0.25B | 如果下一轮融资周期启动时仍没有运营证明,这一情景更可能发生 | 下行由惩罚性条款、优先权堆栈扩大和重复需求疲弱驱动 |
情景区间是分析师假设,不是已报道估值;用途是在明确披露和商业化条件下框定价格纪律。
[CV048, CV049, CV050]公开证据支持宽口径承销区间,而不是 Bose Quantum 的单一公允价值。
情境区间是基于里程碑的承销估计,不应误认为公司披露估值。
[CV048, CV049, CV050]8.5 退出准备度、论点失效触发因素与最终尽调清单
Bose Quantum 还没有公开退出准备度。已保留档案中没有任何信息表明公司具备经审计的公众公司准备度、公开收入透明度,或近期上市安排。更可能的退出是卖给更大的计算、工业或国家相关平台;后续私募融资或资本重组;或者在制造和客户证明改善后,走一条周期很长、逐步扩大披露的路径。这让尽调纪律更重要。论点不应只因量子比特新闻而成立或破裂,而应看决定生存能力的经营事实:可重复付费客户、可信的工厂良率和产出,以及不会把新股权压在隐藏优先权之下的融资条款。给本轮定价之前,投资人应要求经审计收入和毛利率、客户续约分组、股权结构表和分配瀑布细节、制造良率指标,以及可信的出口管制和供应商地图。如果管理层拿不出这些材料,正确动作是继续跟踪,而不是强行给出公开记录无法支撑的精确数字。[CV051, CV052, CV053, CV054]
| 触发因素 | 可观察阈值 | 对投资判断的传导 | 行动含义 |
|---|---|---|---|
| 无重复付费客户 | 管理层无法拿出战略试点之外的续约或扩张证据 | 早期商业化判断失效,部署从收入证明降级成验证秀 | 不为新一轮定价;退回仅跟踪 |
| 工厂有产出但缺少良率证明 | 深圳扩建继续推进,但不披露良率、缺陷率或单位产出指标 | 扩产变成资本消耗,而非护城河扩张 | 加制造风险折价,或直接放弃 |
| 惩罚性融资条款 | 下一轮依赖内部过桥资金、高优先级清算权或结构性保护工具 | 现有融资背书对普通股价值的解释力下降 | 按清算条款重新承销,而不是看名义估值 |
| 出口管制 / 供应商中断 | 关键组件或跨境流程出现许可或采购摩擦 | 执行风险上升,商业化周期可能被拉长 | 缓释措施被证明前,将估值压到区间低端 |
| 叙事跑在经济性前面 | 持续发布量子比特或 AI 消息,但收入披露、客户 cohort 和毛利率仍缺位 | 表明公司仍在证据之前先卖故事 | 维持 research-more,不接受只靠营销做价格发现 |
终止触发因素关注的是存活能力及其如何传导到承销判断,而不只是技术头条里程碑。
[CV014, CV015, CV016, CV020, CV053]| 主题 | 缺失证据 | 为什么重要 | 负责人 / 尽调路径 |
|---|---|---|---|
| 融资条款 | 投后估值、股权类别、清算顺位,以及任何棘轮条款或高优先级保护 | 没有条款,就无法建模回报或下行。 | 向 CFO / 法律顾问索取董事会批准的融资文件和 cap table。 |
| 收入质量 | 经审计收入、毛利率、订单积压和现金消耗 | 收入质量决定公司是在用资金放大增长,还是在补贴概念验证亏损。 | 审阅经审计财报和管理账。 |
| 客户 cohort | 付费客户数量、续约 / 扩张历史,以及按账户计算的集中度 | 区分试点和可投资收入引擎,最干净的证据就在这里。 | 查看 cohort 表和大客户明细。 |
| 制造准备度 | 试点产线良率、单位产出、缺陷率和 capex 路线图 | 没有这些数字,工厂扩张只是资本故事。 | 现场走访,并从 CTO / 制造负责人取得运营资料包。 |
| 政策与供应链 | 出口管制备忘录、供应商地图和国产替代计划 | 跨境摩擦可能同时改写时间线和融资范围。 | 独立贸易律师审查,并访谈供应商。 |
| 治理与退出 | 董事会权利、关联方风险、审计成熟度和长期退出规划 | 即使技术跑通,不透明治理也可能抹掉价值。 | 审阅董事会材料、治理文件和审计准备计划。 |
这些问题是把叙事正面的公司变成可定价投资案例前,至少需要拿到的资料包。
[CV015, CV016, CV051, CV052, CV054]8.6 附录图表
免责声明
本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开;做出任何投资决定前,应直接向管理层和一手文件核验。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Bose Quantum was founded on 2020-11-16 in Beijing. | 高 | SO015, SO016 |
| CO002 | The company's public headquarters is in Chaoyang District, Beijing, at Wanhong West Street. | 高 | SO002, SO015 |
| CO003 | The current legal entity is Beijing Boson Quantum Technology Co., Ltd. as a non-listed joint-stock company. | 高 | SO014, SO015 |
| CO004 | The company positions itself as a full-stack photonic quantum-computing provider spanning hardware, software and cloud access. | 中 | SO001, SO005 |
| CO005 | Public sources consistently describe Bose Quantum as focused on scalable and programmable photonic or coherent optical quantum computing. | 中 | SO003, SO004, SO005 |
| CO006 | Open sources do not disclose audited revenue, ARR, burn or runway for Bose Quantum. | 中 | SO002, SO026 |
| CO007 | Wen Kai is Bose Quantum's founder and CEO. | 高 | SO002, SO005 |
| CO008 | Wen Kai holds a Stanford Ph.D. and has been described by trade press as a former Google quantum lead. | 中 | SO004, SO005, SO017 |
| CO009 | Independent scholarly records show Kai Wen as an active quantum researcher with publications in quantum communication and optical-computing topics. | 高 | SO017, SO018 |
| CO010 | The official team page names Ma Yin, Wei Hai and Wang Chuan as core members of the founding technical leadership bench. | 中 | SO002 |
| CO011 | Aiqicha lists Ma Yin as vice chairman and Li Huiling, Jiang Peixing, Ruan Dong and Wang Dan in financial or director roles. | 中 | SO016 |
| CO012 | The official about page says the core R&D team comes from Stanford, Tsinghua and the Chinese Academy of Sciences. | 中 | SO002 |
| CO013 | Company sources claim that 70 percent of staff are in R&D and 65 percent hold master's or doctoral degrees. | 中 | SO002 |
| CO014 | Boson Quantum had completed seven fundraising events since founding by the time of the late-2024 TMTPost coverage. | 中 | SO005, SO026 |
| CO015 | The March 2026 Series B raised CNY 1 billion, roughly USD 145 million. | 高 | SO007, SO008, SO009, SO010 |
| CO016 | Series B lead investors publicly included Beijing Financial Holdings, ICBC Capital, Chaoyang Shunxi, China Merchants Bank International and Shenzhen Investment Holdings. | 高 | SO007, SO008 |
| CO017 | The 2026 Series B proceeds were earmarked for product R&D, chip-process improvement, a pilot line and manufacturing scale-up, plus quantum-plus-AI commercialization. | 高 | SO007, SO008, SO010 |
| CO018 | The October 2025 A++ round brought in hundreds of millions of renminbi and was co-led by Huade Tech Innovation and Nanshan Strategic Emerging Investment. | 中 | SO006, SO012 |
| CO019 | A++ round proceeds were slated for dedicated and general-purpose coherent photonic quantum computers, chip-fabrication capability and Shenzhen manufacturing operations. | 中 | SO006, SO012 |
| CO020 | A Beijing government-backed high-tech industry fund led Bose Quantum's 2024 A+ financing. | 中 | SO004 |
| CO021 | Bose Quantum's 2023 Series A exceeded RMB 100 million and was tied to China Mobile and Tsinghua-affiliated capital. | 中 | SO013 |
| CO022 | The named investor roster across 2023-2026 rounds indicates that Bose Quantum is financed heavily by state-linked, policy-adjacent or strategic institutional capital. | 中 | SO004, SO006, SO007, SO013 |
| CO023 | Open public evidence does not disclose exact ownership percentages, liquidation preferences, debt obligations or secondary transactions for Bose Quantum's financing history. | 中 | SO014, SO015, SO026 |
| CO024 | Tracxn's accessible output shows Bose Quantum as a Series B company with seven funding rounds, but the post-money valuation field is not openly visible. | 中 | SO026 |
| CO025 | No open source reviewed in this chapter independently verifies the exact post-Series-B valuation often associated with Bose Quantum. | 中 | SO007, SO010, SO026 |
| CO026 | Aiqicha reports Bose Quantum's registered capital at CNY 360 million, which is not the same metric as valuation. | 高 | SO014, SO015 |
| CO027 | Open-source cover metrics leave revenue, ARR, burn and runway unsupported and therefore unsuitable for underwriting from public data alone. | 中 | SO002, SO015, SO026 |
| CO028 | Public English and filing-style sources do not provide enough evidence to verify independent board oversight or full governance rights. | 中 | SO011, SO016 |
| CO029 | Bose Quantum should be analyzed as a private Series B hardware scale-up rather than as a mature public-market issuer or a pre-seed research project. | 中 | SO007, SO015, SO026 |
| CO030 | Bose Quantum publicly launched a 100-qubit coherent photonic quantum-computing system in 2023. | 高 | SO021, SO022 |
| CO031 | By 2024 Bose Quantum had launched a 550-qubit coherent photonic quantum computer and cloud service. | 中 | SO003, SO005 |
| CO032 | The official site says Bose Quantum launched the Yuliang Shanhai 1000 system and its first general photonic quantum chip in May 2026. | 中 | SO002 |
| CO033 | Xinhua reported that the Yuliang Shanhai 1000 can run stably for 7×16 hours and is already being applied in new-drug, materials, brain-science, power and finance scenarios. | 中 | SO007 |
| CO034 | The company claims its 550-qubit cloud platform has recorded over 18 million accesses or solver calls, served more than 400 institutions across 830 disciplines and produced 440-plus application reports. | 中 | SO005 |
| CO035 | Bose Quantum and China Mobile jointly launched the Hengshan photonic quantum-computing platform on the Wuyue cloud platform in December 2023. | 高 | SO021, SO022 |
| CO036 | TMTPost reports that Bose Quantum in 2024 sold China's first commercial photonic quantum computer and issued the first tax invoice for such a system in the country. | 中 | SO004, SO005 |
| CO037 | Official and state-media sources say Bose Quantum is building China's first factory dedicated to large-scale photonic quantum-computer production in Shenzhen, with dozens of units of annual capacity targeted. | 高 | SO019, SO020 |
| CO038 | The official about page says Bose Quantum has built photonic-quantum laboratories in Suzhou, Shenzhen and Nanjing in addition to its Beijing base. | 中 | SO002 |
| CO039 | Aiqicha reports 59 patent records, 114 trademarks and 10 software copyrights for Bose Quantum. | 中 | SO016 |
| CO040 | Physics World's coverage of Baidu and Alibaba exiting direct quantum-computing research is an adverse reminder that commercial quantum hardware remains difficult even in China's strategic sectors. | 中 | SO024 |
| CM001 | Bose Quantum's relevant market is quantum computing systems and services rather than the full umbrella of quantum sensing or quantum communications. | 高 | SM001, SM003, SM012 |
| CM002 | The included spend boundary for Bose covers dedicated photonic hardware, cloud access, developer tooling, and application support rather than generic AI or semiconductor budgets. | 高 | SM001, SM002, SM026 |
| CM003 | For most buyers the status-quo substitutes remain classical HPC, AI models, and classical optimization heuristics, so quantum is entering as a hybrid adjunct rather than a full-stack replacement. | 高 | SM013, SM020, SM023 |
| CM004 | The most repeatedly cited near-term demand pools for quantum computing are optimization, chemistry or materials simulation, and financial modeling. | 高 | SM003, SM012, SM016 |
| CM005 | Bose's photonic positioning is consistent with broader photonic-vendor messaging around modularity, networking, programmability, and easier service delivery than cryogenic lab-only systems. | 中 | SM001, SM017, SM021, SM022 |
| CM006 | QED-C sizes the 2025 global quantum technology market at $1.9 billion and the quantum computing segment at $1.4 billion, with the computing segment rising to more than $3 billion by 2028. | 中 | SM012 |
| CM007 | QED-C counted 7,420 quantum-engaged organizations and 556 pure-play quantum companies by the end of 2025, indicating a broadening supplier and buyer ecosystem before mass deployment. | 中 | SM012 |
| CM008 | McKinsey reports that more than 300 organizations including Airbus, JPMorgan, and Boehringer are already collaborating with quantum technology companies and that first movers are shifting from pilots to workflow-embedded applications. | 高 | SM013, SM024 |
| CM009 | McKinsey says quantum technology startup investment reached $12.6 billion in 2025 while QED-C separately logged $4.9 billion of new private VC and $12.7 billion of new government funding in the same year. | 高 | SM012, SM013 |
| CM010 | QED-C expects the first useful applications to emerge in three to five years, especially in materials science, drug discovery, logistics, and finance. | 高 | SM003, SM012 |
| CM011 | Public quantum market estimates diverge because some sources measure current vendor revenue while others measure broader value pools, investment intensity, or adjacent ecosystem breadth. | 中 | SM012, SM013, SM014 |
| CM012 | Post-Quantum's summary of McKinsey places the 2035 internal quantum technology market at roughly $60 billion to $100 billion overall, with quantum computing accounting for about $43 billion to $71 billion of that range. | 中 | SM014 |
| CM013 | Bose and its China Mobile launch materials show a cloud-first access model in which users start with hosted quantum resources, simulation, and SDK-based task submission before any broader production deployment. | 高 | SM001, SM002, SM026 |
| CM014 | Bose explicitly markets use cases in artificial intelligence, communications, finance, pharmaceuticals, and energy rather than a single vertical wedge. | 高 | SM001, SM002 |
| CM015 | Across photonic vendors, recurring target verticals include chemicals or materials, pharma or drug discovery, finance or optimization, energy, logistics, and cybersecurity. | 高 | SM016, SM020, SM022 |
| CM016 | IQM's 2026 enterprise study says 89 percent of respondents already do hands-on quantum work but only 10 percent report limited production use and 3 percent report scaled deployment. | 中 | SM008 |
| CM017 | IQM reports that about 46 percent of buyers expect on-premises quantum infrastructure to form part of their access model within three years versus 24 percent favoring public cloud alone. | 中 | SM008 |
| CM018 | WJARR argues that QCaaS and hybrid quantum-classical orchestration reduce adoption barriers compared with waiting for fully owned on-prem quantum hardware. | 高 | SM013, SM023 |
| CM019 | Serious buyers increasingly care about calibration access, integration with existing systems, and accumulated organizational capability rather than qubit counts alone. | 高 | SM008, SM013 |
| CM020 | The most plausible early buyer archetypes for Bose are government or HPC institutions, enterprise R&D groups, financial institutions, and industrial optimization teams rather than general IT departments. | 高 | SM003, SM005, SM024, SM025 |
| CM021 | China's quantum ecosystem is state-coordinated and heavily funded, with MERICS estimating about $15 billion of government scientific and industrial spending and USCC highlighting centralized coordination. | 高 | SM003, SM004 |
| CM022 | USCC and TQI both argue that China's state-led structure can accelerate infrastructure buildup while also constraining market-driven experimentation and private-sector autonomy. | 高 | SM003, SM007 |
| CM023 | China says it screened more than 1.3 million university and research-institution patents and identified 680,000 invention patents with commercialization potential, showing a policy push to move science into market channels. | 中 | SM011 |
| CM024 | Chinese official coverage now emphasizes practical quantum use cases such as medical screening, parking or traffic optimization, and drug-related workloads rather than purely laboratory milestones. | 高 | SM009, SM010 |
| CM025 | Public-private partnerships and industrial policy remain core market drivers because QED-C says the industry is still too early-stage and capital-intensive for private capital alone to carry commercialization. | 高 | SM003, SM012 |
| CM026 | Talent is a binding adoption constraint because QED-C says cross-disciplinary quantum talent is scarce and IQM says skills are the most consistent barrier cited by 66 percent or more of large enterprises, universities, and government buyers. | 高 | SM008, SM012 |
| CM027 | Quantum supply chains remain fragile and strategic, with recurring dependencies on photonics, control electronics, cryogenics, and specialized materials. | 高 | SM004, SM012, SM017, SM021 |
| CM028 | Photonic vendors pitch manufacturability, optical-network modularity, and cloud-to-on-prem continuity as adoption advantages over more infrastructure-heavy approaches. | 高 | SM017, SM021, SM022 |
| CM029 | Commercial opportunities remain uncertain even as national-security interest stays high, so China's home market can support infrastructure buildout without proving broad private-market willingness to pay. | 高 | SM003, SM007 |
| CM030 | D-Wave markets production use with paying customers as a point of differentiation, which implies broad market commercialization is still rare enough to be exceptional. | 中 | SM015 |
| CM031 | D-Wave highlights manufacturing, logistics, and life-science deployments, reinforcing that the strongest near-term willingness-to-pay sits where optimization or simulation connects directly to operational metrics. | 高 | SM015, SM016, SM020 |
| CM032 | The photonic market narrative is bifurcated between bold scale claims and practical capability-building, with Bose and peer vendors emphasizing usable systems while adoption studies still show a large production gap. | 中 | SM001, SM008, SM022, SM023 |
| CM033 | The most defensible current sizing frame for Bose is evidence-constrained rather than classic TAM-SAM-SOM because public sources support a low-billions revenue market and an active but still narrow enterprise collaborator base rather than transparent segment conversion data. | 高 | SM003, SM012, SM013 |
| CM034 | The observed buyer journey still runs from education and proofs of concept to hybrid pilot workflows and only then to production or on-prem systems. | 高 | SM008, SM013, SM022, SM023 |
| CM035 | Production photonic vendors typically sell through a layered stack of toolkit or SDK, cloud access, applications support, and then hardware or on-prem systems, which lowers switching costs for early buyers. | 高 | SM002, SM020, SM022, SM026 |
| CM036 | Enterprise quantum benchmarking is already visible in automotive and finance, where public rankings single out companies such as Volkswagen, BMW, JPMorgan, and Goldman Sachs as active quantum adopters. | 中 | SM024, SM025 |
| CM037 | Bose's near-term buyer set is likely to skew toward research-intensive and strategically backed institutions because its public messaging centers on cloud tasks, SDK access, and application pilots rather than off-the-shelf enterprise software. | 高 | SM001, SM002, SM026 |
| CM038 | China-specific commercialization may favor state-linked or public-sector buyers because official reporting stresses strategic platforms, institutional integration, and applied demonstration projects more than transparent competitive pricing. | 高 | SM003, SM004, SM009 |
| CM039 | Even bullish sources frame quantum as a preparation market because QED-C says progress will be gradual and early positioning matters while McKinsey says companies can no longer afford to wait and see. | 高 | SM012, SM013 |
| CM040 | QED-C's current-revenue forecasts and McKinsey's broader value-pool estimates should not be summed because they use different denominators and time horizons. | 高 | SM012, SM014 |
| CM041 | A unit-consistent market pyramid can be expressed in USD billions by anchoring the current market at about 1.4, the low end of the 2035 vendor market at about 60, and the low end of the 2035 economic value lens at about 1300. | 高 | SM012, SM014 |
| CM042 | Bose's practical wedge sits inside the subset of organizations already running pilots or hybrid workflows rather than inside the full theoretical economic value pool. | 高 | SM002, SM008, SM013 |
| CM043 | Buyer attractiveness is strongest where photonic advantages map to existing simulation or optimization workflows and where organizations already fund experimentation. | 高 | SM001, SM016, SM020, SM022 |
| CM044 | The public evidence supports a staged adoption path of learn, cloud experiment, co-develop, integrate, and then scale or install dedicated capacity. | 高 | SM002, SM008, SM022, SM023 |
| CP001 | Bose publicly presents dedicated photonic quantum systems at 100, 550 and 1000 qubits, with an overclocked 1000-series mode that it says can support 3000-qubit problems. | 中 | SP001 |
| CP002 | Bose says its systems run at room temperature without vacuum or ultra-low-temperature infrastructure, which lowers deployment and operations friction relative to cryogenic stacks. | 中 | SP001 |
| CP003 | Bose says its coherent quantum-computing stack has already been integrated into China Mobile Cloud's Wuyue platform for industry applications. | 中 | SP001 |
| CP004 | The retained Bose product and Kaiwu SDK pages show hardware and tooling, but they do not disclose a public list price or standard consumption price. | 高 | SP001, SP002 |
| CP005 | PsiQuantum positions silicon photonics as a modular utility-scale platform that combines photonic chips, networking, cryogenics, control systems and software. | 中 | SP003 |
| CP006 | PsiQuantum publicly emphasizes chemistry, agriculture, energy, defense and PDE-style applications, showing an enterprise-application GTM rather than a transparent self-serve pricing model. | 中 | SP003, SP004 |
| CP007 | ORCA markets PT-1 and PT-2 as rack-mounted, room-temperature photonic systems that can plug into existing HPC environments. | 中 | SP006 |
| CP008 | ORCA says its development environment supports hybrid quantum-classical applications such as combinatorial optimization and generative AI. | 中 | SP007 |
| CP009 | ORCA cites nuclear-fusion modeling and computational-chemistry collaborations, indicating a domain-pilot GTM motion rather than a broad posted-pricing motion. | 中 | SP007 |
| CP010 | Quandela positions MosaiQ as a modular photonic platform available on the cloud and on premises, and says systems from 6 to 24 qubits can be ordered now. | 高 | SP008, SP009 |
| CP011 | Quandela Cloud says it serves hundreds of academic and corporate users, shows 2,461 active users, and offers QPU reservation, APIs and flexible pricing options. | 中 | SP010 |
| CP012 | The retained Xanadu Aurora release supports that Xanadu remains a modular and networked photonic peer, but the retained source does not disclose a public price card or commercialization metrics. | 低 | SP005 |
| CP013 | LightSolver markets a physics-based laser processing unit as an analog alternative for optimization-class HPC workloads and claims roughly 1,000 TOPS-equivalent parallel compute. | 中 | SP011 |
| CP014 | IBM publicly offers a free Open Plan plus PAYG, Flex, Premium and on-prem quantum-compute plans, which is more packaging transparency than Bose or most photonic peers show. | 中 | SP012 |
| CP015 | IBM's retained pricing page lists PAYG at $96 per minute, Flex starting at 400 minutes per year and $72 per minute, Premium starting at 5200 minutes per year and $48 per minute, with on-prem by quote. | 中 | SP012 |
| CP016 | IBM says its Quantum Network includes 300+ members, 65+ commercial partners and startups, 50+ industry clients and 35+ Quantum Innovation Centers. | 中 | SP013 |
| CP017 | D-Wave markets Leap as a real-time cloud service with 99.9% availability, subsecond response times and hybrid solvers. | 中 | SP014 |
| CP018 | D-Wave's 2025 annual report says revenue reached $24.6 million, up 179% year over year. | 中 | SP022 |
| CP019 | The same D-Wave filing says the company has paying customers in production but still needs to accelerate sales cycles and warns that competition can pressure prices and gross margins. | 中 | SP022 |
| CP020 | Amazon Braket exposes a broker model that uses one tooling layer across multiple quantum computers and charges by shot, task or hourly reservation. | 高 | SP015, SP028 |
| CP021 | Amazon Braket publicly highlights access to different hardware modalities including superconducting, trapped-ion and neutral-atom devices. | 高 | SP015, SP028 |
| CP022 | Azure Quantum's public pricing surfaces show that providers set their own prices inside one Microsoft broker layer, and the retained docs list IonQ, Quantinuum, Rigetti and Pasqal offers. | 高 | SP016, SP027 |
| CP023 | The retained Azure docs disclose partner-specific plans including IonQ pay-as-you-go access, Quantinuum subscriptions at $125,000 and $175,000 per month, Rigetti time-based billing, and Pasqal QPU-hour pricing. | 高 | SP016, SP027 |
| CP024 | IonQ says its hardware is integrated with all major cloud platforms, major quantum programming languages and quantum software developer kits. | 中 | SP023 |
| CP025 | IonQ also publicly claims 99.99% two-qubit gate fidelity, 1,200+ patents, five generations of systems in market and deployments on all three major public clouds. | 高 | SP023, SP025 |
| CP026 | Quantinuum says Helios has the highest average two-qubit gate fidelity in the commercial market and showcases enterprise work with BMW, bp, SoftBank and Synopsys. | 中 | SP026 |
| CP027 | Rigetti presents a full-stack chip-to-cloud model and publicly sells the 9-qubit Novera QPU as accessible development hardware. | 中 | SP024 |
| CP028 | NVIDIA argues that useful quantum computing depends on integration with AI supercomputers and hybrid workflows, meaning classical-compute suppliers can capture part of the quantum value chain. | 中 | SP017 |
| CP029 | QED-C says the global 2025 quantum market was about $1.9 billion, with quantum computing at $1.4 billion and forecast above $3 billion by 2028. | 中 | SP021 |
| CP030 | QED-C also says the quantum supply chain is custom and fragile and that quantum talent is still insufficient to support industry growth. | 中 | SP021 |
| CP031 | USCC says state coordination and mega-projects have guided China to early quantum breakthroughs, which supports Bose's domestic-policy alignment story. | 中 | SP019 |
| CP032 | MIT's 2025 report says the United States remains the front-runner in quantum-computing commercialization because of the number and diversity of QPUs, while China leads in quantum communications and patents. | 中 | SP020 |
| CP033 | Physics World reports that Baidu donated its quantum facility to BAQIS and that Alibaba planned to quit direct quantum-computing research, showing commercialization difficulty even for well-capitalized Chinese tech groups. | 中 | SP018 |
| CP034 | Bose's strongest direct overlap is with ORCA, Quandela, Xanadu and PsiQuantum on photonic or optical architectures, so modality alone is not a durable moat. | 中 | SP001, SP003, SP005, SP006, SP008, SP010 |
| CP035 | Bose's clearest defensible wedge in retained public evidence is domestic channel and policy fit rather than uniquely transparent GTM or pricing. | 中 | SP001, SP019, SP020 |
| CP036 | Public multi-vendor clouds from AWS and Azure reduce switching costs by letting buyers test different modalities and providers without committing to a single hardware vendor. | 高 | SP015, SP016, SP027, SP028 |
| CP037 | IBM's network scale and IonQ's all-cloud positioning suggest that distribution power in quantum is concentrating around ecosystem control, not only around qubit-count marketing. | 高 | SP013, SP023, SP025 |
| CP038 | Because Bose does not publicly disclose list pricing, contract structure or paid-customer counts on retained sources, outside buyers can benchmark incumbents more easily than Bose on procurement path and total access cost. | 高 | SP001, SP002, SP012, SP027 |
| CP039 | D-Wave's filing and QED-C together show a category with real commercial traction but still material fragility around long sales cycles, supply chain and talent. | 高 | SP021, SP022 |
| CP040 | LightSolver and NVIDIA indicate that some optimization or hybrid-compute budgets can go to classical or analog accelerators instead of dedicated quantum vendors. | 高 | SP011, SP017 |
| CP041 | Compared with IBM, D-Wave, IonQ, Quantinuum and Rigetti, Bose's public trust signals are narrower because retained Bose pages show technical and channel claims but not audited revenue, broad network counts or global-cloud presence. | 高 | SP001, SP012, SP013, SP022, SP023, SP024, SP026 |
| CP042 | Buyer switching costs into Bose look moderate rather than absolute because SDK and cloud onboarding can be easy, but cloud-broker access makes multi-homing and comparative testing normal. | 高 | SP002, SP015, SP027, SP028 |
| CP043 | Bose uses optimization-oriented public proof points such as a 550-node fully connected Max-Cut result and a 100-variable system claim, which anchors its near-term product story in optimization rather than in general-purpose fault-tolerant computing. | 中 | SP001 |
| CP044 | Western incumbents expose explicit enterprise and government-style procurement paths through on-prem plans, hyperscaler pricing pages and public trust signals, while Bose's retained public materials do not show an equivalent international compliance surface. | 中 | SP012, SP013, SP016, SP019 |
| CI001 | Official surfaces show Bose offers a full stack spanning hardware, cloud platform, SDK, developer community, and vertical solutions rather than a single standalone device. | 中 | SI001, SI003, SI011 |
| CI002 | Bose product pages describe specialized 100-, 550-, and 1,000-qubit systems, with the Shanhai 1000 page claiming support for up to 3,000-bit problems in overclock mode. | 高 | SI003, SI005 |
| CI003 | The homepage lays out a funnel from competitions to community to Kaiwu SDK to cloud validation and then deployment on physical quantum computers. | 中 | SI001 |
| CI004 | China Mobile Hengshan public-beta users can register in the cloud console and order real-machine quantum compute service, showing a usage-based cloud access path. | 高 | SI010, SI013 |
| CI005 | Kaiwu SDK documentation says developers can model QUBO or Ising problems and call Bose real-machine compute directly through the physical interface. | 高 | SI011, SI013 |
| CI006 | The Kaiwu community currently uses free real-machine quotas and certification rewards, including 10 annual 550-qubit quotas and 5 fixed monthly quotas, as a developer-acquisition incentive. | 中 | SI009 |
| CI007 | Bose AI, biopharma, and smart-city solution pages route prospects to consultation rather than checkout, indicating enterprise solution sales instead of self-serve SaaS. | 中 | SI006, SI007, SI008 |
| CI008 | Bose official pages do not publish hardware price cards, per-shot rates, subscription fees, or minimum contract sizes. | 高 | SI001, SI003, SI004 |
| CI009 | That opacity matters because enterprise buyers can benchmark Bose against transparent quantum-access tariffs from IBM, AWS Braket, and Azure Quantum. | 中 | SI022, SI023, SI024 |
| CI010 | IBM publicly lists a free tier plus paid access starting at USD96 per minute for pay-as-you-go, USD72 per minute for flex, and USD48 per minute for premium annual capacity. | 中 | SI022 |
| CI011 | AWS Braket bills quantum-computer use through per-shot plus per-task fees or hourly reservations rather than upfront subscription charges. | 中 | SI023 |
| CI012 | Azure Quantum provider pricing includes explicit execution minimums, with IonQ examples of USD97.50 per program with error mitigation on and USD12.4166 with it off. | 中 | SI024 |
| CI013 | China Mobile describes Bose current cloud service as task-style quantum compute ordered from a console, which points to service or consumption revenue rather than only equipment sales. | 中 | SI013 |
| CI014 | Bose public stack implies at least three revenue lines—hardware systems, cloud or compute access, and vertical solution or integration work—even though the company does not publish revenue mix. | 中 | SI001, SI003, SI013 |
| CI015 | Hardware and solution pages emphasize deployment, reliability, and consultation, which is consistent with quote-based enterprise contracts and longer solution-sales cycles. | 中 | SI003, SI005, SI006 |
| CI016 | Bose has repeatedly used external financing for R&D, chip-process capability, factory buildout, and ecosystem expansion, showing a manufacturing-led scale path rather than a software-light path. | 高 | SI002, SI017, SI019 |
| CI017 | Registry filings show Bose business scope includes cloud equipment sales, cloud equipment manufacturing, instrument manufacturing, and telecom or internet service licenses, confirming exposure to hardware capex and compliance costs. | 高 | SI014, SI015 |
| CI018 | Bose says it has labs in Suzhou, Shenzhen, and Nanjing plus a Shenzhen manufacturing factory, implying ongoing fixed-cost infrastructure beyond headquarters staff. | 中 | SI002 |
| CI019 | Bose argues its systems run at room temperature without vacuum or ultra-low-temperature infrastructure, which should reduce deployment and maintenance burden relative to cryogenic architectures, although the company discloses no margin data. | 中 | SI003, SI021 |
| CI020 | Bose disclosed up to 16+ service hours per day, MTBF of at least 1,000 hours, MTTR of no more than 72 hours, and power draw up to 1,500W for the 1,000-qubit class, implying real uptime, field-service, and energy-cost obligations. | 高 | SI004, SI005 |
| CI021 | The Shanhai 1000 page says automated repair, monitoring, and self-check features are intended to reduce manual O&M cost, but Bose does not publish the resulting service margin. | 高 | SI004, SI005 |
| CI022 | Bose 2026 Series B capital is explicitly earmarked for a pilot quantum-chip production line and factory expansion, meaning commercialization still requires substantial manufacturing investment. | 中 | SI017, SI020, SI021 |
| CI023 | Bose Shenzhen factory began production in November 2025, according to Yicai and follow-on sector coverage. | 中 | SI017, SI020 |
| CI024 | Bose says it has already delivered systems to the National Supercomputing Center in Chengdu, China Mobile Communications Group, Beijing Electronic Zone High-tech Group, and North China University of Technology. | 中 | SI017, SI020 |
| CI025 | The China Mobile platform targets government, enterprise, and research users and exposes 100-qubit real-machine access with authentication, monitoring, and task submission on cloud infrastructure. | 中 | SI013 |
| CI026 | Bose homepage claims a developer community composed of thousands of users, providing a top-of-funnel traction proxy even though conversion rates are undisclosed. | 中 | SI001 |
| CI027 | The Kaiwu Community GitHub repository presents an open Python toolkit with examples for TSP, knapsack, graph coloring, Max-Cut, machine learning, and solver extension, lowering trial friction for developers. | 高 | SI011, SI012 |
| CI028 | Aiqicha company-detail data says Bose had 96 insured employees in 2025, giving a public headcount proxy for current operating scale. | 中 | SI016 |
| CI029 | Bose about-page staffing mix of roughly 70% R&D and 65% masters or PhDs implies a technically heavy payroll base. | 中 | SI002 |
| CI030 | Bose official about page says it has completed multiple financing rounds since founding, aligning with the view that commercial scale-up has depended on recurring outside capital. | 中 | SI002 |
| CI031 | A 2023 Yicai report says Bose raised more than CNY100 million in a round led by the China Mobile Digital New Economy Industry Fund and Tsinghua Holdings Capital. | 中 | SI018 |
| CI032 | A 2025 36Kr report says Bose Series A++ financing of hundreds of millions of yuan was designated for optical-quantum R&D, chip processes, Shenzhen factory construction or operation, and the quantum-plus-AI ecosystem. | 中 | SI019 |
| CI033 | Bose 2026 Series B amounted to CNY1 billion and was led by state-backed and institutional investors for chip-line buildout and factory scale-up. | 中 | SI017, SI020, SI021 |
| CI034 | QCC shows Bose converted into a non-listed joint-stock company in June 2026 and increased registered capital to RMB360 million. | 高 | SI014, SI015 |
| CI035 | Registry pages simultaneously show paid-in capital of about RMB17.98 million, so registered capital should not be treated as a proxy for cash on hand or runway. | 高 | SI014, SI015 |
| CI036 | Public sources in this review do not disclose cash balance, monthly burn, runway, gross margin, ARR, NRR, CAC, payback period, backlog, or customer concentration. | 高 | SI014, SI015, SI017 |
| CI037 | D-Wave 2025 annual-report data show one of the sector most commercial vendors generated USD24.6 million of revenue yet still ended the year with USD884.5 million of cash and marketable securities. | 中 | SI026 |
| CI038 | D-Wave Leap markets 99.9% uptime, subsecond response, and production-grade access, which sets a reliability bar Bose will face when selling cloud quantum services. | 中 | SI025 |
| CI039 | Physics World reported Baidu and Alibaba stepping back from direct quantum-hardware research, reminding investors that long commercialization timelines can still invalidate private-sector quantum bets. | 中 | SI027 |
| CI040 | Because Bose public traction is framed through deployments, platform beta, community quotas, and factory milestones rather than audited revenue, public evidence is insufficient to underwrite revenue quality. | 中 | SI001, SI013, SI017 |
| CI041 | Official solution pages show Bose is pursuing AI, biopharma, energy, finance, communications, and city workloads, which diversifies pipeline targets but also implies integration-heavy, domain-specific selling. | 中 | SI006, SI007, SI008 |
| CI042 | No public source in this review discloses debt facilities, project-finance commitments, or covenant structures for Bose, so absence of evidence should not be mistaken for debt-free operations. | 中 | SI014, SI015, SI017 |
| CI043 | Bose combination of free community quotas and no public production pricing suggests customer acquisition may currently be subsidized to build pipeline, leaving realized pricing and conversion economics opaque. | 中 | SI008, SI009, SI013 |
| CI044 | The public financial verdict is that Bose has credible commercialization surfaces and manufacturing momentum, but the investability case still depends on private disclosure of revenue, margins, conversion, and cash usage rather than on public evidence alone. | 中 | SI002, SI017, SI026 |
| CE001 | Reviewed Bose official surfaces describe a full stack spanning specialized photonic hardware, cloud access, SDK tooling, developer community programs, and vertical solution pages rather than a standalone quantum box. | 中 | SE001, SE008 |
| CE002 | Bose product pages publicly list specialized 100-, 550-, and 1,000-qubit systems, and the flagship Shanhai 1000 page says overclock mode supports up to 3,000-bit problems. | 中 | SE002, SE003, SE004 |
| CE003 | The retained Bose product and SDK materials frame these machines around Ising or QUBO optimization workflows, not around general-purpose gate-based computation. | 中 | SE002, SE010 |
| CE004 | Bose product pages say the systems are fully connected and programmable through a controller, which is central to their optimization-oriented positioning. | 中 | SE002, SE004 |
| CE005 | The public product page describes a dedicated optical-fiber temperature-control module that keeps fiber temperature variation within 0.001°C and adds vibration isolation to protect photonic stability. | 中 | SE002 |
| CE006 | Shanhai 1000 is marketed with standard, precise, and overclock modes plus CAC and QEE error-mitigation features and an L2 intelligent control system. | 中 | SE003, SE004 |
| CE007 | Product specs claim Shanhai 1000 targets at least 16 hours of daily availability, MTBF of at least 1,000 hours, MTTR of at most 72 hours, and power draw of at most 1,500W. | 中 | SE003, SE004 |
| CE008 | The Bose cloud page says its service matrix spans the official cloud platform plus Alibaba Cloud, Huawei Cloud, Tencent Cloud, and China Mobile Cloud channels. | 中 | SE008 |
| CE009 | The cloud page says Quantum Cloud Hub virtualizes 24 dedicated photonic machines into one resource pool with dynamic load balancing and automatic failover. | 中 | SE008 |
| CE010 | Bose publicly presents four access modes: direct cloud service, major public-cloud integration, a fusion-compute route, and local hardware deployment. | 中 | SE008 |
| CE011 | Kaiwu SDK documentation describes a Python environment for solving QUBO problems on coherent photonic quantum computers with core modules including qubo, cim, ising, preprocess, classical, sampler, solver, and common. | 中 | SE010 |
| CE012 | Kaiwu installation docs require Python 3.10, Bose platform credentials, and a local wheel install rather than a simple public package-registry-only flow. | 中 | SE011 |
| CE013 | The Kaiwu changelog shows the product maturing from matrix import and simulators in v0.9.0 to QuboModel, direct solver support, model-to-real-machine tutorials, CIMOptimizer, PrecisionReducer, and automatic license handling by v1.3.0. | 中 | SE012 |
| CE014 | The real-machine tutorial documents two public execution paths: upload a QUBO matrix in the cloud console or invoke CIMOptimizer and PrecisionReducer directly from the SDK. | 中 | SE014, SE010 |
| CE015 | The TSP tutorial shows Bose teaching the same optimization problem across classical SA solving and CPQC-style real-machine execution, which makes the SDK look like a workflow bridge rather than a standalone cloud form. | 中 | SE013, SE014 |
| CE016 | The qboson/kaiwu_community repository presents a pip-installable toolkit for QUBO work, custom solver extension, and example-driven use cases such as TSP and Max-Cut. | 中 | SE015 |
| CE017 | The Kaiwu-PyTorch-Plugin repository claims PyTorch-native support for RBM, BM, QVAE, and QDiffusion workflows while retaining access to photonic quantum sampling through Kaiwu backends. | 中 | SE016, SE005 |
| CE018 | The kaiwu-basic-examples repository includes notebooks and scripts for QUBO matrices, TSP, Max Cut, and a CIM optimizer example, giving the community a practical onboarding path beyond slides. | 中 | SE017, SE014 |
| CE019 | Developer-harbor organizes community code into AI, biomedical research, and communication categories, which broadens the developer surface but still looks like a curated code showcase rather than a mature marketplace. | 中 | SE018 |
| CE020 | The Kaiwu community portal offers 550-qubit quotas and community certification badges, showing Bose uses subsidized real-machine access as a developer-conversion tactic. | 中 | SE009, SE017 |
| CE021 | The AI solution page positions QBM and Kaiwu-PyTorch-Plugin as the public AI layer, emphasizing energy-based training, full-connectivity structure, and model-acceleration use cases. | 中 | SE005, SE016 |
| CE022 | The biopharmaceutical page extends the same stack into sequence design, docking, cell-state modeling, and materials discovery, indicating Bose sells domain workflows rather than a separate life-science instrument. | 中 | SE006, SE016 |
| CE023 | The city solution page maps the platform to bus-route planning, MIMO beamforming, anti-fraud, and energy optimization, reinforcing that public use cases are optimization-heavy. | 中 | SE007, SE013 |
| CE024 | The Quantum and arXiv paper coauthored by Kai Wen describes coherent Ising machines built from antisymmetrically coupled degenerate optical parametric oscillator pulses plus dissipative pulses and nonlinear transfer functions. | 中 | SE019, SE020 |
| CE025 | The broader photonic-computing review treats photonics as a mainstream path for scalable and potentially room-temperature-friendly architectures, which is consistent with Bose’s modality choice but not a direct validation of Bose execution. | 中 | SE021, SE001 |
| CE026 | Jiemian reports that Bose compares its special-purpose machines to a "quantum GPU," explicitly betting on narrower systems that can reach the market before universal machines. | 中 | SE022 |
| CE027 | Jiemian and Quantum Computing Report both describe Bose as focused on special-purpose photonic optimization tasks rather than general-purpose fault-tolerant computing. | 中 | SE022, SE024 |
| CE028 | Jiemian reports access through China Mobile Cloud, Alibaba Cloud, and Huawei Cloud, plus deployment at Chengdu Supercomputing Center, which gives independent support for at least some institutional integration. | 中 | SE022 |
| CE029 | Yicai says the 2026 Series B will fund a chip pilot line, factory expansion, and an AI-plus-quantum ecosystem, linking manufacturing scale-up directly to the product roadmap. | 中 | SE023, SE024 |
| CE030 | Quantum Computing Report says the Shenzhen factory began operations in November 2025 and is meant to standardize manufacturing and improve hardware reliability. | 中 | SE024, SE023 |
| CE031 | SCIO says the Shenzhen facility will contain module development, full-system production and manufacturing, and quality-control-and-testing divisions. | 中 | SE025 |
| CE032 | SCIO says the facility targets production of dozens of photonic quantum computers annually, which is a more explicit manufacturing-capacity claim than the main Bose website offers. | 中 | SE025, SE024 |
| CE033 | The Bose homepage roadmap moves from 100-, 550-, and 1,000-qubit generations toward transformer-based tuning, chaotic amplitude control, multi-core parallelism, and a CQ-H photonic-chip program. | 中 | SE001 |
| CE034 | The best-publicly-evidenced near-term differentiation is in control-system packaging and reliability messaging—modes, smart control, uptime, and failover—rather than in a broad independent benchmark corpus. | 中 | SE003, SE004, SE008 |
| CE035 | The only explicit third-party validation visible in the reviewed Bose product specs is a company claim of China Academy of Information and Communications Technology technical verification, without a certificate link on the retained page. | 低 | SE004 |
| CE036 | IAF CertSearch is designed to verify active certificates by company name or certificate number, but no Bose-specific security or quality certificate artifact appeared in the reviewed materials for this chapter. | 中 | SE026, SE001 |
| CE037 | Bose official pages market enterprise-grade security, data isolation, and cloud usability, but the reviewed public surfaces did not yield a trust center, privacy/security whitepaper, SOC report, ISO certificate, or public incident archive. | 中 | SE001, SE008, SE026 |
| CE038 | Jiemian includes adverse outside views that special-purpose quantum systems may remain commercially narrow and could face pressure from improving classical AI hardware. | 中 | SE022 |
| CE039 | The chapter’s strongest technical-moat evidence comes from room-temperature photonic operation, full connectivity, and PyTorch-compatible tooling, not from a large public patent list or broad package-registry distribution. | 中 | SE001, SE002, SE016 |
| CE040 | The main product-technology diligence blocker is that Bose exposes workflow and uptime claims publicly but not the independent artifacts needed to verify security controls, SLA performance, or chip-process yield. | 中 | SE004, SE008, SE023, SE026 |
| CE041 | A 2023 China Mobile / Bose article says the Hengshan public beta integrated Bose's 100-qubit machine into a one-stop flow covering data construction, task submission, secure authentication, status monitoring, and task-style ordering from a console. | 中 | SE027 |
| CU001 | Retained Bose-owned customer surfaces consistently target biopharma, finance, AI, transport or city operations, materials, and adjacent industrial users rather than a single generic buyer persona. | 中 | SU001, SU002, SU003, SU004, SU026 |
| CU002 | The official cloud page says Bose has already cooperated with more than 50 universities and enterprises. | 中 | SU001 |
| CU003 | Bose presents four commercial access modes: direct cloud service, mainstream cloud integration, fused compute deployment, and local hardware deployment. | 中 | SU001 |
| CU004 | The cloud page publishes task-based list pricing of 128 RMB per run for SPQC-1000 and 68 RMB per run for SPQC-550. | 中 | SU001 |
| CU005 | The official cloud surface explicitly supports private deployment for buyers that purchase specialized quantum hardware. | 中 | SU001 |
| CU006 | Bose says its cloud control plane manages 24 machines as one pooled resource with dynamic load balancing and automatic failover. | 中 | SU001 |
| CU007 | Kaiwu materials show an open-source community edition for QUBO work and a separate enterprise flow that sends modeled jobs to real quantum hardware, which is the clearest retained separation between free tooling and paid machine access. | 中 | SU016, SU024, SU025 |
| CU008 | Jiemian reports that users can access Bose machines through China Mobile Cloud, Alibaba Cloud, and Huawei Cloud on a pay-per-use basis. | 中 | SU005 |
| CU009 | QbitAI says the China Mobile Hengshan photonic-quantum platform launched in public beta on 1 December 2023. | 中 | SU006 |
| CU010 | The same China Mobile launch coverage says the service is open to government, enterprise, and research users after registration inside the Wuyue quantum cloud console. | 中 | SU006 |
| CU011 | The China Mobile launch report says the platform exposes a 100-computational-qubit machine, Kaiwu SDK access, and simulation services as a one-stop task-submission workflow. | 中 | SU006 |
| CU012 | The 2023 China Mobile launch article lists prior Bose collaborations with China Mobile, Ping An Bank, Huaxia Bank, Qianfang Tech, the Beijing Academy of Quantum Information Sciences, and the Chinese Academy of Medical Sciences. | 低 | SU006 |
| CU013 | Jiemian says about 39 percent of Bose users come from biopharmaceuticals, including work with Guangzhou National Laboratory, XtalPi, and BGI. | 中 | SU005 |
| CU014 | Jiemian says universities account for much of the remaining usage, which implies the visible user mix is still research heavy. | 中 | SU005 |
| CU015 | Xinhua says the National Supercomputing Center in Chengdu has deployed a 550-qubit coherent photonic machine and integrated it with 100P FP64 classical compute. | 高 | SU007, SU008 |
| CU016 | Xinhua says the Chengdu supercomputing platform has already completed task tests with a fintech research team, which is stronger than logo-only proof but still short of a disclosed recurring contract. | 中 | SU007 |
| CU017 | The Sohu coverage says the Chengdu project includes a unified scheduling cloud platform and joint work with University of Electronic Science and Technology of China and Southwest Minzu University. | 低 | SU008 |
| CU018 | Securities Times says Bose won China Merchants Bank’s first quantum-computing procurement project, Tiancheng AI, in October 2025. | 中 | SU011 |
| CU019 | The Securities Times report says the China Merchants Bank project includes task-based real-machine service, optimization interfaces, fund-pool data validation, and custom algorithm support. | 中 | SU011 |
| CU020 | Science and Technology Daily says Bose already has deep cooperation with China Mobile, XtalPi, and Shenzhen Metro and claims those projects have validated solution quality and efficiency. | 中 | SU021 |
| CU021 | Science and Technology Daily says Bose has already landed specialized quantum use cases in biopharma, brain science, and new materials, with some problems solving more than 80 percent faster. | 中 | SU020 |
| CU022 | QbitAI says cumulative solve calls have exceeded 68 million since the 100-qubit cloud service launched. | 中 | SU012 |
| CU023 | The same QbitAI report says the platform covers more than 900 institutions and more than 10,000 participating developers. | 中 | SU012 |
| CU024 | QbitAI says customers have already used Bose-trained QBM-VAE workflows in peptide generation, small-molecule generation, single-cell clustering, mRNA vaccine design optimization, and proteomics analysis. | 中 | SU012 |
| CU025 | QbitAI says Bose has active exploration relationships with Guangzhou National Laboratory, Shanghai Jiao Tong University, Sun Yat-sen University’s pharmacy school, Beijing Cancer Hospital, and Tsinghua Changgeng Hospital. | 中 | SU012 |
| CU026 | The Kaiwu developer community page says the community includes articles, Q&A, learning modules, events, and direct login into the real-machine cloud platform. | 中 | SU023 |
| CU027 | Saikr shows Bose using the 2026 MathorCup modeling competition for mass developer education and lead generation. | 中 | SU017 |
| CU028 | The public GitHub organization page shows only two active flagship repos with visible recent updates, which indicates a real but still narrow open-source surface relative to the size of Bose’s claimed commercial ecosystem. | 低 | SU015 |
| CU029 | The Kaiwu Community repo is organized around QUBO modeling, example code, issue reporting, and community contributions, reinforcing a developer-led top-of-funnel. | 中 | SU013, SU016 |
| CU030 | The Kaiwu PyTorch plugin targets machine-learning developers by wrapping Boltzmann-model training and evaluation inside a familiar PyTorch workflow. | 中 | SU014, SU025 |
| CU031 | Tencent News quotes Bose co-founder Ma Yin saying banks and insurers remain cautious about high-risk, long-payback hard-tech projects and often require stronger revenue or financing profiles. | 中 | SU019 |
| CU032 | Jiemian says supercomputing centers are Bose’s first customers because they offer public funding, technical validation, and lower commercial risk. | 中 | SU005 |
| CU033 | Jiemian also says rival engineers view such supercomputing-center projects as exploratory and more aligned with local-government expectations than with sustained commercial demand. | 中 | SU005 |
| CU034 | Jiemian says most Chinese quantum firms remain loss-making and reliant on venture funding rather than repeat customers. | 中 | SU005 |
| CU035 | None of the retained public customer sources disclose NRR, GRR, logo churn, renewal rate, contract length, or satisfaction scores, so durability is still unproven externally. | 中 | SU001, SU005, SU011, SU012, SU023 |
| CU036 | The retained public record clearly separates broad usage surfaces from paid deployment proof only in a few cases: China Mobile cloud subscriptions, Chengdu supercomputing deployment, and the China Merchants Bank procurement win. | 中 | SU006, SU007, SU011 |
| CU037 | Because Bose publishes per-task cloud pricing and describes the bank engagement as task-based real-machine service, the visible monetization surface is usage-led rather than seat-led. | 中 | SU001, SU011 |
| CU038 | Xinhuanet frames commercialization as the core test for China’s “quantum plus AI” wave, reinforcing that the sector is still judged by application transfer rather than by disclosed renewal metrics. | 中 | SU022 |
| CU039 | Zhejiang Online shows Ping An Bank Hangzhou branch conducting onsite visits to Bose and peers, which evidences financial-sector interest but not yet a disclosed production contract. | 中 | SU018 |
| CU040 | The QBoson about page says the company has completed scenario validation across AI, biopharma, finance, communications, energy, new materials, and transport. | 中 | SU026 |
| CU041 | Science and Technology Daily says Bose signed strategic agreements with more than 20 partners at its 2026 launch event, which suggests ecosystem expansion even though customer and revenue splits remain undisclosed. | 中 | SU020 |
| CU042 | Jiemian cites researchers who question whether customers will keep paying for dedicated quantum hardware as classical alternatives continue to improve. | 中 | SU005 |
| CR001 | The U.S. Treasury's outbound-investment program explicitly covers China-linked quantum information technologies as a national-security category. | 高 | SR001, SR006, SR008 |
| CR002 | The final outbound-investment rule effective January 2, 2025 adds notification, prohibition, diligence, and record-keeping burdens for covered China-quantum transactions. | 高 | SR001, SR006, SR009, SR029 |
| CR003 | BIS guidance says a license is required for advanced-computing items shipped to entities headquartered in Country Group D:5, which can constrain a China-headquartered quantum hardware company even when parts or collaborations route through third countries. | 高 | SR002, SR030 |
| CR004 | China's dual-use export-control regulations effective December 1, 2024 apply to goods, technology, services, and data with civil-military or proliferation relevance, creating a domestic transfer-and-export compliance layer around quantum systems. | 高 | SR003, SR004 |
| CR005 | Lawfare's review of May 2025 controls says 37 Chinese entities were added to the Entity List, including 22 connected to quantum advances, showing that China quantum activity is already a live target of tightening U.S. controls. | 中 | SR005, SR010 |
| CR006 | Bird & Bird and MERICS both describe China quantum development as strategically prioritized and heavily state-guided rather than broadly market-led. | 中 | SR004, SR011 |
| CR007 | Qichacha shows QBoson's disclosed business scope includes internet information services and first-class value-added telecom services that depend on domestic approvals. | 中 | SR024 |
| CR008 | Qichacha shows the company converted to a non-listed joint-stock company and raised registered capital to RMB 360 million in June 2026. | 中 | SR024 |
| CR009 | The retained public sources surface patent activity and legal-scope changes, but they do not by themselves establish a verified no-litigation or no-sanctions conclusion across every company name variant. | 低 | SR024, SR025 |
| CR010 | Yicai says QBoson plans to use its Series B proceeds for technological breakthroughs, a quantum-chip pilot line, and expansion of China's first large-scale quantum-computer factory. | 中 | SR020, SR021 |
| CR011 | Quantum Computing Report says the Shenzhen factory effort is intended to standardize manufacturing processes and improve hardware reliability, not just add capacity. | 中 | SR020, SR021 |
| CR012 | QBoson says it is simultaneously operating labs in Suzhou, Shenzhen, and Nanjing while building manufacturing capability in Shenzhen, which increases coordination burden. | 中 | SR025 |
| CR013 | Jiemian says order intake outpaced hand-built prototypes, pushing QBoson toward factory construction before the economics of universal quantum computing are settled. | 中 | SR022 |
| CR014 | The coherent-Ising review says repeated digital-to-analog and analog-to-digital conversions are a bottleneck that prevents optics from realizing its full speed-and-power advantage without major PIC advances. | 中 | SR014 |
| CR015 | The single-photon CIM paper frames hybrid analog-plus-digital control as a necessary trade-off between physics speed and classical precision rather than a solved scaling pattern. | 中 | SR015 |
| CR016 | The photonic-quantum review says scalable, fault-tolerant photonic quantum computing remains a frontier challenge despite rapid architectural progress. | 中 | SR016 |
| CR017 | PostQuantum's photonic ecosystem map identifies photon sources, superconducting nanowire detectors, optical switching, and photonic integrated circuits as separate bottlenecks, which means QBoson cannot solve scale by mastering only one layer. | 中 | SR018, SR016 |
| CR018 | PostQuantum's China supply-chain analysis says local progress is real, but severe vulnerabilities still remain across the full stack and the 80% localization narrative requires skepticism. | 中 | SR017 |
| CR019 | The Quantum Insider reports that quantum systems are exposed to chokepoints in niobium, nickel, rare earths, and other specialized materials where China holds direct or indirect leverage. | 中 | SR019 |
| CR020 | CNAS argues that the next three to five years are the industrial transition from laboratory systems to deployable quantum infrastructure, so supply chains and integration now gate commercialization as much as science does. | 中 | SR013, SR012 |
| CR021 | Jiemian says users can access QBoson machines through China Mobile Cloud, Alibaba Cloud, and Huawei Cloud on a pay-per-use basis, making distribution partners central to reach and monetization. | 中 | SR022 |
| CR022 | QbitAI says the China Mobile launch wraps data construction, task submission, authentication, monitoring, and message transfer into one partner-cloud workflow, embedding QBoson into a larger platform stack. | 中 | SR023 |
| CR023 | Jiemian says about 39% of disclosed users come from biopharmaceuticals while universities account for much of the remaining mix, implying visible demand is not yet diversified across commercial verticals. | 中 | SR022 |
| CR024 | Jiemian says supercomputing centers are low-risk first customers because they offer public funding and validation, which means early deployments can be more exploratory than revenue-durable. | 中 | SR022 |
| CR025 | Quantum Computing Report names Chengdu supercomputing, China Mobile, and North China University of Technology as visible deployments, supporting customer proof but also revealing concentration in a small set of institutional anchors. | 中 | SR020 |
| CR026 | MERICS says private capital and companies play a smaller role in China's quantum ecosystem than in the U.S., with firms often acting as intermediaries between state priorities and research institutions. | 中 | SR011 |
| CR027 | USCC says China's quantum progress is rapid but uneven and state coordination drives much of the system, which raises execution risk if policy resources shift or civilian demand lags. | 中 | SR010, SR011 |
| CR028 | Yicai shows QBoson is deploying RMB 1 billion into scale-up before public revenue disclosure, which is a direct sign of capital intensity ahead of proven self-funding commercialization. | 中 | SR021 |
| CR029 | Jiemian says revenue figures are undisclosed and industry insiders still characterize most Chinese quantum firms as loss-making and venture-funded rather than repeat-customer funded. | 中 | SR022 |
| CR030 | D-Wave's 2025 10-K says the company expects additional operating losses and may need capital sooner than planned, illustrating that public quantum hardware vendors still finance through sustained losses. | 中 | SR026 |
| CR031 | Rigetti's 2025 10-K says a significant percentage of revenue depends on a limited number of customers and on public-sector contracts. | 中 | SR028 |
| CR032 | Quantum Computing Inc.'s 2024 10-K says substantial doubt existed about continuing as a going concern without additional capital and that several technologies remained early in commercialization. | 中 | SR027 |
| CR033 | Across D-Wave, Quantum Computing Inc., and Rigetti, public filings show that listed quantum vendors still combine heavy R&D burn, delayed profitability, and concentration risk even with broader disclosure and capital-market access than QBoson. | 高 | SR026, SR027, SR028 |
| CR034 | QBoson says roughly 70% of staff are in R&D and that it has applied for dozens of invention patents, which provides some technical-depth mitigation against pure research risk. | 中 | SR025 |
| CR035 | Qichacha lists 96 insured employees in the 2025 annual report, which is meaningful for a young startup but still modest relative to a chip line, factory, cloud stack, and multi-city labs. | 中 | SR024, SR025 |
| CR036 | Qichacha records board and personnel changes in June 2026 while the company was changing legal form and expanding capital, which raises governance-transition risk during scale-up. | 中 | SR024 |
| CR037 | RAND's framework says commercialization depends on scientific research, government support, private industry activity, and technical achievement together, so weakness in one leg can slow the whole industrial base. | 中 | SR012, SR013 |
| CR038 | Treasury and multiple law-firm analyses all describe the outbound-investment regime as expansive and diligence-heavy, so compliance friction should be treated as durable rather than symbolic. | 高 | SR001, SR006, SR007, SR008, SR009, SR029 |
| CR039 | Bird & Bird, MERICS, and USCC together imply that state support can speed scale-up while also biasing early traction toward policy-friendly institutions instead of broad market demand. | 中 | SR004, SR010, SR011 |
| CR040 | Qichacha and QbitAI together show that regulated cloud-service and partner-authenticated access are already part of QBoson's commercial surface, so domestic compliance outages could interrupt actual service delivery, not just paperwork. | 中 | SR024, SR023 |
| CR041 | The highest residual risk is commercialization mismatch: factory and chip-line scale can advance faster than repeat paying demand, leaving QBoson with more fixed cost before unit economics are proven. | 中 | SR020, SR021, SR022, SR026, SR027, SR028 |
| CR042 | Export-control friction, supply-chain chokepoints, and partner concentration can compound into delayed deployments, weaker revenue visibility, and another financing cycle. | 中 | SR002, SR003, SR017, SR018, SR019, SR020 |
| CR043 | No retained public source provides an uptime history, SLA archive, recall record, or factory-yield disclosure for QBoson's production and cloud stack. | 低 | SR020, SR022, SR023, SR025 |
| CR044 | No retained public source discloses revenue mix, renewal rates, top-customer share, or channel split, so concentration cannot be underwritten from public evidence alone. | 低 | SR020, SR022, SR025 |
| CR045 | No retained public source discloses a supplier list or BOM-level dependence for detectors, photon sources, PIC fabrication, or EDA/tool-chain inputs. | 低 | SR017, SR018, SR025 |
| CV001 | Bose Quantum says it is focused on practical special-purpose quantum computing rather than a general-purpose universal stack. | 中 | SV001, SV002 |
| CV002 | Bose Quantum publicly presents hardware, cloud, and application-solution surfaces rather than only a research prototype. | 中 | SV001, SV002 |
| CV003 | QBoson completed a Series B round worth CNY 1 billion (about $145 million) in March 2026. | 中 | SV003, SV004, SV005, SV006 |
| CV004 | The 2026 Series B investor roster included Beijing Financial Holdings, ICBC Capital, CMB International, Shenzhen Investment Holdings, and Addor Capital. | 中 | SV003, SV004 |
| CV005 | Public reporting says the Series B proceeds will fund a pilot chip line, Shenzhen factory expansion, and technical scale-up. | 中 | SV003, SV005, SV006 |
| CV006 | Boson Quantum publicly disclosed an A+ round in 2024 led by a Beijing government-backed industry fund. | 中 | SV007 |
| CV007 | Boson Quantum disclosed a follow-on A++ or second-phase Series A+ round in October 2025. | 中 | SV008, SV010 |
| CV008 | TMTPost described the 2025 financing as the company’s seventh disclosed fundraising event. | 中 | SV008 |
| CV009 | Boson Quantum said it sold China’s first commercial photonic quantum computer and issued the first tax invoice for such a system in 2024. | 中 | SV008 |
| CV010 | Boson Quantum’s 550-qubit cloud platform had been accessed more than 18 million times by late 2025, according to TMTPost. | 中 | SV008 |
| CV011 | Jiemian reported that Bose Quantum has not publicly disclosed revenue figures. | 中 | SV009 |
| CV012 | Jiemian said industry insiders describe most Chinese quantum companies as loss-making and dependent on venture funding rather than repeat customers. | 中 | SV009 |
| CV013 | Jiemian said Bose Quantum’s earliest customers are often supercomputing centers that provide validation and public funding but not broad commercial diversity. | 中 | SV009 |
| CV014 | Jiemian said Bose Quantum had begun building a factory in Shenzhen by late 2025. | 中 | SV009, SV005 |
| CV015 | The retained public funding sources disclose amounts and investors but do not disclose Bose Quantum’s post-money valuation, security type, or preference terms. | 中 | SV003, SV004, SV005, SV006, SV007, SV008, SV010 |
| CV016 | The retained Bose Quantum sources do not disclose revenue, gross margin, burn, runway, or customer-retention cohorts. | 中 | SV003, SV004, SV009 |
| CV017 | USCC says China’s quantum push is shaped by state coordination, mega-project funding, and security priorities rather than only market pull. | 中 | SV011 |
| CV018 | MERICS says China’s quantum development is state-governed and increasingly oriented toward supply-chain localization. | 中 | SV012 |
| CV019 | RAND says significant private-sector commercialization in quantum is still new and eventual applications remain highly uncertain. | 中 | SV013 |
| CV020 | Lawfare says export-control regimes and civil-military-fusion concerns broaden risk around China-linked quantum activity. | 中 | SV014 |
| CV021 | IonQ reported $130.0 million of annual revenue for fiscal 2025 and said it was the first quantum company above $100 million of GAAP annual revenue. | 中 | SV015 |
| CV022 | IonQ reported $3.3 billion of cash, cash equivalents, and investments as of December 31, 2025. | 中 | SV015 |
| CV023 | IonQ’s market capitalization was about $19.88 billion around June 30, 2026 according to market-data trackers. | 中 | SV016, SV017 |
| CV024 | D-Wave reported $24.6 million of fiscal 2025 revenue, up 179% from $8.8 million in fiscal 2024. | 中 | SV018 |
| CV025 | D-Wave said it ended 2025 with its highest liquidity position in company history at more than $884 million. | 中 | SV018 |
| CV026 | D-Wave’s market capitalization was about $8.88 billion around June 30, 2026 according to market-data trackers. | 中 | SV019, SV020 |
| CV027 | Rigetti reported $7.1 million of total revenue for fiscal 2025. | 中 | SV021, SV022 |
| CV028 | Rigetti reported a GAAP net loss of $216.2 million for fiscal 2025. | 中 | SV021 |
| CV029 | Rigetti reported $589.8 million of cash, cash equivalents, and available-for-sale investments as of December 31, 2025. | 中 | SV021 |
| CV030 | Rigetti’s market capitalization was about $6.42 billion around June 30, 2026, while its 2025 10-K recorded a $3.82 billion non-affiliate market value at June 30, 2025. | 中 | SV022, SV023, SV024 |
| CV031 | Quantum Computing Inc. reported $373 thousand of total revenue for fiscal 2024 in its 2024 Form 10-K. | 中 | SV025 |
| CV032 | Stock Analysis said Quantum Computing Inc. generated $682 thousand of annual revenue in 2025 and $4.33 million of trailing-twelve-month revenue by March 31, 2026. | 中 | SV027 |
| CV033 | Quantum Computing Inc. had a market capitalization of about $2.19 billion around June 30, 2026 according to market-data trackers. | 中 | SV026, SV028 |
| CV034 | Using market capitalization around the run date and the latest disclosed annual revenue, IonQ trades near 152.9x revenue. | 低 | SV015, SV016, SV017 |
| CV035 | Using market capitalization around the run date and the latest disclosed annual revenue, D-Wave trades near 361.0x revenue. | 低 | SV018, SV019, SV020 |
| CV036 | Using market capitalization around the run date and the latest disclosed annual revenue, Rigetti trades near 904.2x revenue. | 低 | SV021, SV022, SV023, SV024 |
| CV037 | Using the 2025 revenue series from Stock Analysis and the June 2026 market cap snapshot, Quantum Computing Inc. trades near 3211.1x revenue. | 低 | SV027, SV028 |
| CV038 | Quantinuum announced a roughly $600 million capital raise at a $10 billion pre-money valuation in September 2025. | 中 | SV029 |
| CV039 | ORCA Computing announced a $15 million Series A round in 2022 to commercialize photonic quantum systems. | 中 | SV030 |
| CV040 | The jump from ORCA’s early-stage photonic financing to Quantinuum’s scaled $10 billion mark shows that private quantum valuation dispersion is too wide to transfer mechanically to Bose Quantum. | 低 | SV029, SV030 |
| CV041 | Public pure-play quantum valuations are pricing long-duration optionality, strategic scarcity, and cash runway rather than mature unit economics. | 低 | SV015, SV018, SV021, SV027 |
| CV042 | Because Bose Quantum has disclosed fundraising but not revenue, post-money valuation, or preference terms, public evidence cannot support a buy recommendation at an undisclosed price. | 中 | SV003, SV004, SV009, SV015, SV018, SV021 |
| CV043 | If Bose Quantum were marketed at a premium close to the richer public-quantum optionality set, investors would be paying for factory scale and repeat demand before the public evidence proves either. | 低 | SV005, SV009, SV015, SV018, SV021 |
| CV044 | The most defensible current recommendation is research-more rather than buy or avoid. | 中 | SV003, SV009, SV015, SV018, SV021, SV029 |
| CV045 | Confidence in that recommendation is medium because the product and financing signals are real but the financial disclosure gap is material. | 中 | SV003, SV008, SV009, SV015, SV018 |
| CV046 | The risk rating is high because commercialization risk, policy risk, and financing opacity reinforce one another. | 中 | SV009, SV011, SV012, SV013, SV014 |
| CV047 | The cleanest valuation stance is unknown on disclosed facts alone, although any aggressive markup over the latest round would look stretched without private proof on revenue and renewal quality. | 中 | SV003, SV009, SV015, SV018, SV021 |
| CV048 | A bull case would require repeat enterprise or government renewals, disclosed revenue scale, pilot-line proof, and manageable policy constraints before a roughly $0.8–1.2 billion value band is defensible. | 低 | SV005, SV009, SV029, SV030 |
| CV049 | A base case that assumes pilot-heavy commercialization and another opaque financing step supports a broader $0.35–0.6 billion underwriting band. | 低 | SV009, SV015, SV018, SV021 |
| CV050 | A bear case where demand or policy execution breaks and the next round becomes a bridge or down-round supports only a roughly $0.1–0.25 billion value band. | 低 | SV009, SV013, SV014, SV030 |
| CV051 | Bose Quantum is not IPO-ready on public evidence because audited revenue disclosure, term-sheet transparency, and mature governance signals are absent from the retained sources. | 中 | SV003, SV009, SV025 |
| CV052 | The most plausible exit paths today are a strategic sale, a later state-linked recapitalization, or another private round rather than a near-term public listing. | 低 | SV009, SV029, SV030 |
| CV053 | The key thesis-break triggers are failure to show repeat paying customers, failure to show factory yield and unit output, or a financing event with punitive terms. | 中 | SV009, SV013, SV014 |
| CV054 | A pricing decision should be gated on audited revenue and gross margin, customer-renewal cohorts, cap-table and liquidation terms, factory-yield metrics, and an export-control memo. | 中 | SV009, SV014, SV025 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | QBoson | 玻色量子 - 专用量子计算机 | 国际领先的量子计算技术 | 玻色量子聚焦相干量子计算及光量子计算技术路线,致力于可扩展、可编程光量子计算的各类型软硬件全平台研发与产业落地。 |
| SO002 | QBoson | 玻色量子 - 关于玻色量子 | 北京玻色量子科技股份有限公司于2020年底在北京市朝阳区成立,是一家专注于专用量子计算的硬科技公司。 |
| SO003 | Baidu Baike | Beijing Boson Quantum Technology Co., Ltd. | Beijing Municipality Boson Quantum Technology Co., Ltd. was founded in November 2020 and is headquartered in Chaoyang District, Beijing. |
| SO004 | TMTPost | Beijing Government-backed Fund Leads A+ Round Financing for Boson Quantum | The funding round was led by Beijing High-Precision and Cutting-edge Industry Development Investment Fund. |
| SO005 | TMTPost | China's Boson Quantum Secures New Funding to Accelerate Photonic Quantum Computing Push | Boson Quantum in 2024 sold China's first commercial photonic quantum computer and issued the first official tax invoice for such a system. |
| SO006 | TMTPost | QBoson Secures Multi-Hundred-Million-USD Series A++ Funding to Advance Photonic Quantum Computing and AI Integration | The funding round was co-led by Huade Tech Innovation and Nanshan Strategic Emerging Investment. |
| SO007 | Xinhua | 玻色量子完成10亿元B轮融资,加速专用量子计算产业化落地 | 3月31日,北京玻色量子科技有限公司完成10亿元B轮融资。 |
| SO008 | Yicai Global | Chinese Quantum Startup QBoson Bags USD145 Million to Expand Chip, Computer Production | QBoson has completed a Series B funding round worth CNY1 billion (USD145 million). |
| SO009 | Quantum Computing Report | QBoson Secures CNY 1 Billion ($145 Million USD) Series B for Photonic Quantum Hardware Scaling | QBoson, a Beijing-based developer of photonic quantum hardware, has closed a Series B funding round totaling CNY 1 billion ($145 million). |
| SO010 | The Quantum Insider | Chinese Quantum Startup QBoson Raises $145 Million to Scale Chip Production | QBoson has raised CNY 1 billion — or about $145 million USD — to expand chip production and scale its quantum computing systems. |
| SO011 | The Quantum Insider | China's Quantum Sector Sees Investment Surge as Larger Funding Rounds Return | Beijing-based photonic quantum company QBoson closed a CNY 1 billion ($145 million) Series B. |
| SO012 | 36Kr | 玻色量子完成数亿元A++轮融资,诺奖开启量子计算大航海时代 | 本轮融资由华德科创、南山战新投联合领投,广发信德、湖南财信产业基金、纬德信息等跟投。 |
| SO013 | Yicai Global | Chinese Quantum Computing Startup QBoson Raises Over USD14.4 Million | Chinese Quantum Computing Startup QBoson Raises Over USD14.4 Million. |
| SO014 | QCC / Qichacha | 北京玻色量子科技股份有限公司 | 2026-06-07 企业名称变更:从北京玻色量子科技有限公司变更为北京玻色量子科技股份有限公司。 |
| SO015 | Aiqicha | 北京玻色量子科技股份有限公司 - 工商信息查询 - 爱企查 | 注册时间:2020-11-16;注册资本:36,000万(元);企业类型:其他股份有限公司(非上市)。 |
| SO016 | Aiqicha | 北京玻色量子科技股份有限公司 - 玻色量子 - 爱企查 | 参保人数为96人,其中马寅担任副董事长,李惠玲担任财务负责人。 |
| SO017 | Google Scholar | Kai Wen | Kai Wen's profile lists quantum-computing and quantum-communication publications across Stanford and other research collaborations. |
| SO018 | Quantum Journal | Combinatorial optimization solving by coherent Ising machines based on spiking neural networks | The paper credits Kai Wen and Chuan Wang in work linking coherent Ising machines and optical spiking neural networks. |
| SO019 | State Council Information Office / Xinhua | China's first photonic quantum computer factory breaks ground in Shenzhen | Once completed, the facility is expected to manufacture dozens of photonic quantum computers annually. |
| SO020 | People's Daily | China's first photonic quantum computer factory breaks ground in Shenzhen | Compared to other technological paths, photonic quantum computing does not require ultra-low temperatures. |
| SO021 | China.com.cn | 中国移动联合玻色量子打造“人人可用的量子计算”——恒山光量子算力平台公测上线 | 中国移动云能力中心联合北京玻色量子科技有限公司共同打造的五岳量子计算云平台——恒山光量子算力平台正式开启公测。 |
| SO022 | QBoson Official CSDN Blog | 中国移动联合玻色量子打造“人人可用的量子计算”——恒山光量子算力平台公测上线 | 这是玻色量子继2023年5月16日成功发布自研的国内首台100量子比特相干光量子计算机真机之后的又一重要里程碑。 |
| SO023 | Fortune China | 2025年《财富》中国科技50强 - 北京玻色量子科技有限公司 | 玻色量子在今年发布的1000量子比特相干光量子计算机成为中国首个达到该级别的光量子计算设备。 |
| SO024 | Physics World | Baidu and Alibaba plan to quit quantum computing research | Baidu and Alibaba have announced they are quitting quantum research, an adverse signal on near-term commercial viability. |
| SO025 | GitHub | qboson organization repositories | The qboson organization publicly lists Kaiwu-PyTorch-Plugin and Kaiwu Community repositories with recent 2026 updates. |
| SO026 | Tracxn | Bose Quantum funding and investors | The accessible Tracxn output shows 7 funding rounds and a Series B, but masks post-money valuation details behind subscription access. |
| SM001 | QBoson | 玻色量子产品系列 - 量子计算机硬件解决方案 | |
| SM002 | China Internet News Center | 中国移动联合玻色量子打造“人人可用的量子计算”——恒山光量子算力平台公测上线_中国网 | |
| SM003 | U.S.-China Economic and Security Review Commission | Vying for Quantum Supremacy: U.S.-China Competition in Quantum Technologies | U.S.- CHINA | Central direction and a subordinate role for the commercial sector may constrain market-driven innovation in China’s quantum development. |
| SM004 | MERICS | China’s long view on quantum tech has the US and EU playing catch-up | China sees quantum technology as pivotal in global science and technology competition and has stepped up government spending on scientific and industrial development to about USD 15 billion. |
| SM005 | MIT Initiative on the Digital Economy | Recently Released: 2025 MIT Quantum Index Report - MIT Initiative on the Digital Economy | |
| SM006 | MIT Initiative on the Digital Economy | MIT Report Describes Quantum Opportunities, Challenges - MIT Initiative on the Digital Economy | |
| SM007 | The Quantum Insider | State vs. Market: The Pros And Cons of China And U.S.'s Quantum Innovation Models | |
| SM008 | IQM Quantum Computers | New Industry Study Finds Quantum Computing Has Entered a Capability Era, With Early Movers Building an Advantage Later Entrants Will Struggle to Close - IQM Quantum Computers | Enterprise engagement is now nearly universal but production use remains rare: 89% of respondents report hands-on quantum work, while only 10% report limited production use and 3% have reached scaled deployment. |
| SM009 | State Council Information Office | China races to turn quantum computing into industrial solutions | |
| SM010 | The State Council of the People's Republic of China | China hits new landmark in global quantum computing race | |
| SM011 | The State Council of the People's Republic of China | China completes patent screening at universities, research institutions to enhance commercialization | |
| SM012 | QED-C | State of the Global Quantum Industry 2026 | QED-C | The quantum computing market is scaling rapidly from a $1.4B market to more than $3B over the same period. |
| SM013 | McKinsey & Company | McKinsey Quantum Technology Monitor 2026: A commercial tipping point | Over 300 organizations including Airbus, Boehringer Ingelheim, E.ON, JPMorgan Chase, and Liberty Mutual are actively collaborating with quantum technology companies to solve business challenges. |
| SM014 | Post-Quantum | McKinsey Quantum Monitor 2026: Tipping Point? | |
| SM015 | D-Wave Quantum Inc. | D-Wave Quantum Inc. Annual Report 2025 | We are delivering measurable results today – applications in production, with paying customers, solving real-world problems that classical systems cannot solve. |
| SM016 | PsiQuantum | Applications — PsiQuantum | |
| SM017 | PsiQuantum | Technology — PsiQuantum | |
| SM018 | Xanadu | Xanadu | Xanadu introduces Aurora: world's first scalable, networked and modular quantum computer | |
| SM019 | Xanadu | Xanadu | Welcome to Xanadu | |
| SM020 | ORCA Computing | Applications | ORCA Computing | |
| SM021 | ORCA Computing | Technology | ORCA Computing | |
| SM022 | Quandela | Quandela | Leading Photonic Quantum Computing Solutions | |
| SM023 | World Journal of Advanced Research and Reviews | Quantum cloud computing: Enterprise strategies for hybrid quantum-classical workloads | |
| SM024 | The Quantum Insider | Tracking Quantum Adoption in Enterprise: The Quantum Insider and HorizonX Consulting Launch Quantum Index at IYQ | |
| SM025 | Quantum Innovation Index | Quantum Innovation Index | |
| SM026 | QBoson | Welcome to Kaiwu SDK — Kaiwu SDK 1.3.1 documentation | |
| SP001 | QBoson | 玻色量子产品系列 - 量子计算机硬件解决方案 | 目前已经成功研制出可实用化的100、550、1000和3000(超频模式下)计算量子比特的四款专用量子计算机。 |
| SP002 | QBoson | Welcome to Kaiwu SDK — Kaiwu SDK 1.3.1 documentation | |
| SP003 | PsiQuantum | Technology — PsiQuantum | PsiQuantum's modular architecture combines photonic chips, networking, cryogenics, control systems, and software into an integrated platform designed for scalable deployment. |
| SP004 | PsiQuantum | Applications — PsiQuantum | |
| SP005 | Xanadu | Xanadu introduces Aurora: world's first scalable, networked and modular quantum computer | Xanadu introduces Aurora: world's first scalable, networked and modular quantum computer. |
| SP006 | ORCA Computing | Technology | ORCA Computing | Rack-mounted, room temperature and leveraging telecoms-grade optical fibre components, the PT Series is an ideal solution for organisations seeking to integrate quantum computing into existing HPC infrastructure without special consideration for quantum. |
| SP007 | ORCA Computing | Applications | ORCA Computing | |
| SP008 | Quandela | Quandela | Leading Photonic Quantum Computing Solutions | Photonic quantum computers: modular, scalable, energy-efficient, accessible on the cloud and on-premises. |
| SP009 | Quandela | Products And Services | Order quantum computers from 6 up to 24 qubits now. |
| SP010 | Quandela | Cloud | Initially launched in November 2022, and now adopted by hundreds of academic and corporate users, Quandela Cloud is a comprehensive platform to discover, learn, test and develop quantum solutions. |
| SP011 | LightSolver | Technology - LightSolver | The Laser Processing Unit™ introduces physics-based computing, enabling efficient, analog processing of complex HPC workloads that are computationally intensive on traditional hardware. |
| SP012 | IBM | IBM Quantum Computing | Products and services | Pay-As-You-Go Plan — Pay for the quantum computer access you use. Billed per second of usage. |
| SP013 | IBM | IBM Quantum Computing | IBM Quantum Network | Meet the world’s largest network of quantum innovators. |
| SP014 | D-Wave | The Leap™ Quantum Cloud Service | D-Wave | The Leap™ quantum cloud service provides real-time access to the world’s largest quantum computers and powerful hybrid solvers while offering production-grade accessibility, reliability, and security. |
| SP015 | Amazon Web Services | Amazon Braket Pricing | Amazon Braket offers three pricing components for on-demand use of a quantum computer (QPU): a per-shot fee and a per-task fee or a single hourly reservation fee. |
| SP016 | Microsoft Azure | Azure Quantum - Pricing | Microsoft Azure | Azure Quantum offers a range of quantum solutions offered by Microsoft and its partners. Microsoft partners define and control the pricing of the solutions they make available. |
| SP017 | NVIDIA | Quantum Computing Solutions from NVIDIA | Turning QPUs into useful quantum computers means integrating them with state-of-the-art AI supercomputers. |
| SP018 | Physics World | Baidu and Alibaba plan to quit quantum computing research | The Chinese search engine company Baidu is giving up its quantum computing division by donating its entire research facility to the government-run Beijing Academy of Quantum Information Sciences (BAQIS). |
| SP019 | U.S.-China Economic and Security Review Commission | Vying for Quantum Supremacy: U.S.-China Competition in Quantum Technologies | State Coordination Guides China to Early Quantum Breakthroughs. |
| SP020 | MIT Initiative on the Digital Economy | MIT Report Describes Quantum Opportunities, Challenges | By contrast, the United States is the clear leader in quantum computing research quality and impact. The country is foremost in the number and diversity of quantum processing units (QPUs), making it the front-runner in quantum computing’s commercialization. |
| SP021 | QED-C | State of the Global Quantum Industry 2026 | Quantum computing is scaling rapidly from a $1.4B market to more than $3B over the same period. |
| SP022 | Securities and Exchange Commission / D-Wave Quantum Inc. | D-Wave Quantum Inc. Annual Report 2025 | In 2025, we delivered record revenue of $24.6 million, up 179% year-over-year. |
| SP023 | IonQ | IonQ - Investor Relations | We lead the market with the first and only quantum computing hardware integrated with all major cloud platforms, quantum programming languages, and quantum software developer kits. |
| SP024 | Rigetti | Quantum Computing | The Novera QPU, our 9-qubit QPU, gives you unprecedented access to quantum technology and empowers you to take your research to the next level. |
| SP025 | IonQ | IonQ: Trapped Ion Quantum Computing Company | 99.99% two-qubit gate fidelity. |
| SP026 | Quantinuum | Quantinuum | Accelerating Quantum Computing | Helios has the highest average two-qubit gate fidelity of any commercial quantum computer in the industry. |
| SP027 | Microsoft Learn | Pricing Plans for Azure Quantum Providers - Azure Quantum | In Azure Quantum, hardware and software providers define and control the pricing of their offerings. |
| SP028 | Amazon Web Services | Cloud Quantum Computing Service - Amazon Braket - AWS | Easily work with different types of quantum computers and circuit simulators using a consistent set of development tools. |
| SI001 | QBoson | 玻色量子 - 专业的量子计算解决方案提供商 | 从开发者大赛激发创意灵感,加入活跃的开发者社区,通过直观的Kaiwu SDK构建算法模型,然后在量子云平台验证性能,最终部署至实体量子计算机解决真实世界挑战。 |
| SI002 | QBoson | 关于玻色量子 - 专业的量子计算公司 | 公司已在苏州、深圳、南京等地建设光量子实验室,并于深圳落地中国首个规模化专用量子计算机制造工厂。 |
| SI003 | QBoson | 玻色量子产品系列 - 量子计算机硬件解决方案 | 相干量子计算已完成工程样机和算法验证,已并入中国移动云“五岳量子计算云平台”,可在人工智能、通信、金融、医药、能源等行业实际应用。 |
| SI004 | QBoson | 玻色量子产品规格对比 - 详细技术参数 | 驭量山海1000 可用时长不少于16小时/天,MTBF不少于1000小时,功耗不超过1,500W。 |
| SI005 | QBoson | 驭量·山海1000专用量子计算机 | 整体服务时长16+小时/天;L2级自动调光控制、常见故障智能修复,减少人工运维成本。 |
| SI006 | QBoson | AI 人工智能解决方案 - 量子计算赋能智能应用 | |
| SI007 | QBoson | 生物制药解决方案 - 量子计算加速新药研发 | |
| SI008 | QBoson | 智慧城市解决方案 - 量子计算驱动城市创新 | |
| SI009 | Kaiwu Quantum Developer Community | 开物量子开发者社区 | Pass Rewards 10 quotas for 550-qubit real quantum machines with a one-year validity period; Fixed Monthly Benefits: 5 quotas for 550-qubit real quantum machines. |
| SI010 | QBoson | 相干光量子计算云平台-玻色量子 | |
| SI011 | QBoson | Overview — Kaiwu SDK 1.3.1 documentation | The SDK is a software development kit designed for solving QUBO problems ... and call the quantum computer machine directly through the physical interface. |
| SI012 | GitHub | GitHub - qboson/kaiwu_community | |
| SI013 | China Internet Information Center | 中国移动联合玻色量子打造“人人可用的量子计算”——恒山光量子算力平台公测上线 | 用户在注册开通“五岳”量子云服务后,进入控制台页面,选择访问“恒山光量子算力服务”,即可订购真机算力服务。 |
| SI014 | Qichacha | 北京玻色量子科技股份有限公司 - 企查查 | 注册资本 36000万元;实缴资本 1798.0521万元;参保人数 96 (2025年报)。 |
| SI015 | Aiqicha | 北京玻色量子科技股份有限公司 - 工商信息查询 - 爱企查 | |
| SI016 | Aiqicha | 北京玻色量子科技股份有限公司 - 玻色量子 - 爱企查 | |
| SI017 | Yicai Global | Chinese Quantum Startup QBoson Bags USD145 Million to Expand Chip, Computer Production | The proceeds will be used to ... establish a pilot production line for quantum computing chips, expand operations at China’s first large-scale quantum computer factory, and build a business ecosystem integrating quantum computing with artificial intelligence. |
| SI018 | Yicai Global | Chinese Quantum Computing Startup QBoson Raises Over USD14.4 Million | |
| SI019 | 36Kr | Bose Quantum Secures Hundreds of Millions of Yuan in Series A++ Financing | The funds will be continuously used for ... construction of quantum computing chip process capabilities; construction and operation of the first large-scale special-purpose optical quantum computer manufacturing factory in China in Nanshan District, Shenzhen. |
| SI020 | The Quantum Insider | Chinese Quantum Startup QBoson Raises $145 Million to Scale Chip Production | |
| SI021 | Quantum Computing Report | QBoson Secures CNY 1 Billion ($145 Million USD) Series B for Photonic Quantum Hardware Scaling | |
| SI022 | IBM Quantum | IBM Quantum Computing | Products and services | |
| SI023 | Amazon Web Services | Amazon Braket Pricing | |
| SI024 | Microsoft Learn | Pricing Plans for Azure Quantum Providers - Azure Quantum | |
| SI025 | D-Wave | The Leap™ Quantum Cloud Service | D-Wave | |
| SI026 | U.S. Securities and Exchange Commission / D-Wave Quantum | D-Wave Quantum Inc. Annual Report 2025 | In 2025, we delivered record revenue of $24.6 million, up 179% year-over-year, and we ended the year with $884.5 million in cash and marketable securities. |
| SI027 | Physics World | Baidu and Alibaba plan to quit quantum computing research | Given it takes many years before quantum products fully hit the market, the shift in focus to other business activities may be commercially driven. |
| SE001 | QBoson | 玻色量子 - 专业的量子计算解决方案提供商 | 从开发者大赛激发创意灵感,加入活跃的开发者社区,通过直观的Kaiwu SDK构建算法模型,然后在量子云平台验证性能,最终部署至实体量子计算机解决真实世界挑战。 |
| SE002 | QBoson | 玻色量子产品系列 - 量子计算机硬件解决方案 | 通过控制器可以实现任意节点之间的全连接,可软件编程适配不同场景。 |
| SE003 | QBoson | 玻色量子 - 专用量子计算机 | 国际领先的量子计算技术 | 国内首款最高支持3000比特且可长时间稳定运行的专用量子计算机,支持3种运算模式。 |
| SE004 | QBoson | 玻色量子产品规格对比 - 详细技术参数 | 驭量山海1000可用时长不少于16小时/天,MTBF不少于1000小时,MTTR不超过72小时;通过中国信息通信研究院技术验证。 |
| SE005 | QBoson | AI 人工智能解决方案 - 量子计算赋能智能应用 | |
| SE006 | QBoson | 生物制药解决方案 - 量子计算加速新药研发 | |
| SE007 | QBoson | 智慧城市解决方案 - 量子计算驱动城市创新 | |
| SE008 | QBoson Cloud | 玻色量子云 | 量子真机算力云服务 | 多台专用光量子计算机虚拟化为统一算力池,24台设备统一纳管、动态负载均衡、故障自动切换。 |
| SE009 | Kaiwu Quantum Developer Community | 开物量子开发者社区-聚焦十五五规划,深耕专用实用化量子计算学习实践,玻色量子实操与开发者认证平台 | Pass Rewards 10 quotas for 550-qubit real quantum machines with a one-year validity period; Fixed Monthly Benefits: 5 quotas for 550-qubit real quantum machines. |
| SE010 | QBoson | Overview — Kaiwu SDK 1.3.1 documentation | The SDK currently includes the following core modules: qubo, cim, ising, preprocess, classical, sampler, solver, and common. |
| SE011 | QBoson | Installation Instructions — Kaiwu SDK 1.3.1 documentation | |
| SE012 | QBoson | Changelog — Kaiwu SDK 1.3.1 documentation | |
| SE013 | QBoson | Beginner Tutorial - QUBO Modeling | |
| SE014 | QBoson | Beginner Tutorial - Using the Real Machine | The whole process from modeling, submitting Qubo matrix to the cloud platform, and obtaining calculation results from the cloud platform is demonstrated through the Traveling Salesman Problem (TSP). |
| SE015 | GitHub / qboson | GitHub - qboson/kaiwu_community | |
| SE016 | GitHub / qboson | GitHub - qboson/kaiwu-pytorch-plugin | Kaiwu-PyTorch-Plugin is a quantum computing programming suite based on PyTorch and the Kaiwu SDK. |
| SE017 | GitHub / QBosonCommunity | GitHub - QBosonCommunity/kaiwu-basic-examples | |
| SE018 | GitHub / QBosonCommunity | GitHub - QBosonCommunity/Developer-harbor | |
| SE019 | Quantum | Combinatorial optimization solving by coherent Ising machines based on spiking neural networks | |
| SE020 | arXiv | Combinatorial optimization solving by coherent Ising machines based on spiking neural networks | |
| SE021 | arXiv | Photonic Quantum Computers | |
| SE022 | Jiemian Global | Selling before perfection: a Chinese startup tests a shortcut to quantum computing-Jiemian Global | Not everyone is convinced. Researchers working on universal quantum systems argue that special-purpose machines face limits of their own. |
| SE023 | Yicai Global | Chinese Quantum Startup QBoson Bags USD145 Million to Expand Chip, Computer Production | The proceeds will be used to overcome key technological barriers to practical quantum computers, establish a pilot production line for quantum computing chips, expand operations at China’s first large-scale quantum computer factory, and build a business ecosystem integrating quantum computing with artificial intelligence. |
| SE024 | Quantum Computing Report | QBoson Secures CNY 1 Billion ($145 Million USD) Series B for Photonic Quantum Hardware Scaling | |
| SE025 | State Council Information Office / Xinhua | China's first photonic quantum computer factory breaks ground in Shenzhen | |
| SE026 | IAF CertSearch | Verify Certificates Instantly | IAF CertSearch | A certificate is valid if it appears in IAF CertSearch, has an Active status, and is up to date. |
| SE027 | China Internet Information Center | 中国移动联合玻色量子打造“人人可用的量子计算”——恒山光量子算力平台公测上线_中国网 | 端到端实现“数据构建、任务提交、安全鉴权、状态监控、消息互传”的一站式支持,对外提供持续稳定的、任务式的量子真机算力服务。 |
| SU001 | QBoson | 玻色量子云 | 量子真机算力云服务 | 玻色量子已与50+家高校及企业合作并获得众多领域专家支持。 |
| SU002 | QBoson | AI 人工智能解决方案 - 量子计算赋能智能应用 | |
| SU003 | QBoson | 生物制药解决方案 - 量子计算加速新药研发 | |
| SU004 | QBoson | 智慧城市解决方案 - 量子计算驱动城市创新 | |
| SU005 | Jiemian Global | Selling before perfection: a Chinese startup tests a shortcut to quantum computing-Jiemian Global | Commercial traction remains tentative. |
| SU006 | QbitAI | 人人可用的量子计算!恒山光量子算力平台公测上线 | “恒山光量子算力平台”面向政企及科研用户开放。 |
| SU007 | Xinhua / news.cn | “量子+超算”融合创新平台落地成都 | 国家超算成都中心已完成550量子比特相干光量子计算机部署。 |
| SU008 | Sohu | 玻色量子“量超融合”再突破:国内首台专用量子计算机部署国家超算中心 | |
| SU011 | Securities Times / STCN | 玻色量子中标招商银行量子计算采购项目“天秤AI” | 公司中标招商银行首个量子计算采购项目“天秤AI”。 |
| SU012 | QbitAI | 量子+AI4S!玻色量子完成数亿A++轮融资 | 平台调用求解次数累计超过6800w次,覆盖院校超过900所,参与研发的开发者人数超过10000人。 |
| SU013 | GitHub | GitHub - qboson/kaiwu_community | |
| SU014 | GitHub | GitHub - qboson/kaiwu-pytorch-plugin | |
| SU015 | GitHub | qboson organization repositories | |
| SU016 | Read the Docs | 概述 — Kaiwu Community 1.0.4 文档 | |
| SU017 | Saikr | 4月8日(今晚)19:30,玻色量子企业专家公开讲座开课通知-2026年第十六届MathorCup数学应用挑战赛-赛氪 | |
| SU018 | Zhejiang Online / ZJOL | 以"金融之芯"托举"量子强国" ——平安银行杭州分行开展量子产业专题调研 | |
| SU019 | Tencent News / Beijing Business Today | 玻色量子创始人&COO马寅:银行在投资高风险、长周期回报的硬科技项目时较为谨慎 | 银行投资机构在面对高风险、长周期回报的硬科技项目时显得较为谨慎。 |
| SU020 | Science and Technology Daily | 齐聚第二十八届科博会 玻色量子构建量子计算产业全新生态 | 现场,玻色量子与20余家合作伙伴签署战略协议。 |
| SU021 | Science and Technology Daily | 我国首个规模化专用光量子计算机制造工厂落地深圳 | 目前,公司已与中国移动、晶泰科技、深圳地铁等开展深度合作。 |
| SU022 | Xinhua / Xinhuanet | 新华网财经观察丨融资热潮来袭 “量子+AI”加速产业化落地 | |
| SU023 | 量科网 / Quantum Technology Center | 玻色量子助力国内量子计算技术实用化 开物量子开发者社区正式上线! | |
| SU024 | QBoson | Kaiwu SDK社区版GitHub仓库地址 | |
| SU025 | QBoson | Welcome to Kaiwu SDK — kaiwu SDK Documentation v1.1.2 documentation | |
| SU026 | QBoson | 关于玻色量子 - 专业的量子计算公司 | |
| SU027 | QBoson | 玻色量子产品系列 - 量子计算机硬件解决方案 | |
| SR001 | U.S. Department of the Treasury | Outbound Investment Security Program | |
| SR002 | Bureau of Industry and Security | Guidance Regarding Enforcement of License Requirements for Advanced Computing Items for Entities Headquartered in Country Group D:5 and Macau | |
| SR003 | State Council of the People's Republic of China | Regulations of the People's Republic of China on Export Control of Dual-Use Items | |
| SR004 | Bird & Bird | Quantum Computing Laws and Regulations 2026 – China | |
| SR005 | Lawfare | Technology Controls to Contain China's Quantum Ambitions Are Here | |
| SR006 | Kirkland & Ellis | U.S. Department of the Treasury Releases Final Rule Implementing Executive Order on Outbound Foreign Investments into China | |
| SR007 | Gibson Dunn | Much Ado About Outbound: Unpacking the Newest U.S. Regulatory Regime | |
| SR008 | Goodwin | Twenty Questions on the Outbound Investment Security Program | |
| SR009 | Skadden | US Treasury Creates the Reverse CFIUS Program, a Limited Great Wall on Outbound Investment | |
| SR010 | U.S.-China Economic and Security Review Commission | Vying for Quantum Supremacy: U.S.-China Competition in Quantum Technologies | |
| SR011 | MERICS | China's long view on quantum tech has the US and EU playing catch-up | |
| SR012 | RAND | An Assessment of the U.S. and Chinese Industrial Bases in Quantum Technology | |
| SR013 | Center for a New American Security | Quantum's Industrial Moment | |
| SR014 | arXiv | Coherent Ising Machines: The Good, The Bad, The Ugly | |
| SR015 | IOP Quantum Science and Technology | Single photon coherent Ising machines for constrained optimization problems | |
| SR016 | arXiv | Photonic Quantum Computers | |
| SR017 | PostQuantum | China's Quantum Supply Chain: How Export Controls Are Building What They Sought to Prevent | |
| SR018 | PostQuantum | The Fab's Hidden Supply Chain: Who Really Wins If Photonic Quantum Computing Wins | |
| SR019 | The Quantum Insider | Scientists Propose Early-Warning System for Quantum Supply Chain as China Tightens Minerals Grip | |
| SR020 | Quantum Computing Report | QBoson Secures CNY 1 Billion ($145 Million USD) Series B for Photonic Quantum Hardware Scaling | |
| SR021 | Yicai Global | Chinese Quantum Startup QBoson Bags USD145 Million to Expand Chip, Computer Production | |
| SR022 | Jiemian Global | Selling before perfection: a Chinese startup tests a shortcut to quantum computing | |
| SR023 | QbitAI | Hengshan Photonic Quantum Computing Platform Public Beta Launch | |
| SR024 | Qichacha | Beijing Boson Quantum Technology Co., Ltd. company record | |
| SR025 | QBoson | About QBoson | |
| SR026 | D-Wave Quantum | Annual report on Form 10-K for fiscal year 2025 | |
| SR027 | Quantum Computing Inc. | Annual report on Form 10-K for fiscal year 2024 | |
| SR028 | Rigetti Computing | Annual report on Form 10-K for fiscal year 2025 | |
| SR029 | Davis Polk | Final outbound investment rule released | |
| SR030 | Bureau of Industry and Security | Export Administration Regulations landing page | |
| SV001 | QBoson | 玻色量子 - 专业的量子计算解决方案提供商 | 聚焦专用量子计算和量子计算产业化。 |
| SV002 | QBoson | 关于玻色量子 - 专业的量子计算公司 | |
| SV003 | Yicai Global | Chinese Quantum Startup QBoson Bags USD145 Million to Expand Chip, Computer Production | |
| SV004 | Quantum Computing Report | QBoson Secures CNY 1 Billion ($145 Million USD) Series B for Photonic Quantum Hardware Scaling | |
| SV005 | The Quantum Insider | Chinese Quantum Startup QBoson Raises $145 Million to Scale Chip Production | |
| SV006 | Xinhua | 玻色量子完成10亿元B轮融资,加速专用量子计算产业化落地 | |
| SV007 | TMTPost | Beijing Government-backed Fund Leads A+ Round Financing for Boson Quantum | |
| SV008 | TMTPost | China's Boson Quantum Secures New Funding to Accelerate Photonic Quantum Computing Push | |
| SV009 | Jiemian Global | Selling before perfection: a Chinese startup tests a shortcut to quantum computing-Jiemian Global | Revenue figures have not been disclosed. Industry insiders say most Chinese quantum firms remain loss-making, relying on venture funding rather than repeat customers. |
| SV010 | 36Kr | 玻色量子完成数亿元A++轮融资, 诺奖开启量子计算大航海时代 | |
| SV011 | U.S.-China Economic and Security Review Commission | Vying for Quantum Supremacy: U.S.-China Competition in Quantum Technologies | |
| SV012 | MERICS | China’s long view on quantum tech has the US and EU playing catch-up | |
| SV013 | RAND | An Assessment of the U.S. and Chinese Industrial Bases in Quantum Technology | |
| SV014 | Lawfare | Technology Controls to Contain China’s Quantum Ambitions Are Here | |
| SV015 | IonQ | IonQ Announces Fourth Quarter and Full Year 2025 Financial Results | |
| SV016 | CompaniesMarketCap | IonQ (IONQ) - Market capitalization | |
| SV017 | Macrotrends | IonQ Market Cap 2021-2025 | IONQ | |
| SV018 | D-Wave Quantum | D-Wave Reports Fourth Quarter and Year-End 2025 Results | |
| SV019 | CompaniesMarketCap | D-Wave Quantum (QBTS) - Market capitalization | |
| SV020 | Macrotrends | D-Wave Quantum Market Cap 2021-2025 | QBTS | |
| SV021 | Rigetti | Rigetti Computing Reports Fourth Quarter and Full-Year 2025 Financial Results | |
| SV022 | Securities and Exchange Commission | RIGETTI COMPUTING, INC. Annual Report for fiscal year ended December 31, 2025 | |
| SV023 | CompaniesMarketCap | Rigetti Computing (RGTI) - Market capitalization | |
| SV024 | Macrotrends | Rigetti Computing Market Cap 2021-2025 | RGTI | |
| SV025 | Securities and Exchange Commission | Quantum Computing Inc. Annual Report for fiscal year ended December 31, 2024 | |
| SV026 | CompaniesMarketCap | Quantum Computing (QUBT) - Market capitalization | |
| SV027 | Stock Analysis | Quantum Computing (QUBT) Revenue 2011-2026 | |
| SV028 | Stock Analysis | Quantum Computing (QUBT) Market Cap & Net Worth | |
| SV029 | Quantinuum / Honeywell | Honeywell Announces $600 Million Capital Raise For Quantinuum at $10b Pre-Money Equity Valuation to Advance Quantum Computing at Scale | |
| SV030 | ORCA Computing | ORCA Computing completes $15 million Series A funding round | ORCA Computing |